GORMAN - 2010 - Trading Zones Interactional Expertise

SCIENCE, TECHNOLOGY, AND SOCIETY
TRADING ZONES AND INTERACTIONAL EXPERTISE
TRADING ZONES
CREATING NEW KINDS OF COLLABORATION
edited by Michael E. Gorman
Cross-disciplinary collaboration increasingly characterizes today’s science and engineering research.

The problems and opportunities facing society do not come neatly sorted by discipline. Difficulties
arise when researchers from disciplines as different as engineering and the humanities work together
and find that they speak largely different languages. This book explores a new framework for fostering
collaborations among existing disciplines and expertise communities. The framework unites two ideas
that have emerged from recent work in STS: trading zones, in which scientific subcultures, each with its
AND
own language, develop the equivalents of pidgin and creole; and interactional expertise, in which experts
learn to use the language of another research community in ways that are indistinguishable from expert
INTERACTIONAL EXPERTISE
practitioners of that community. A trading zone can gradually become a new area of expertise, facilitated
by interactional expertise and involving negotiations over boundary objects (objects represented in
different ways by different participants).
The volume describes applications of the framework to service science, business strategy,
environmental management, education, and practical ethics. One detailed case study focuses
on attempts to create trading zones that would help prevent marine bycatch; another investigates
trading zones formed to market the female condom to women in Africa; another describes how
humanists embedded in a nanotechnology laboratory gained interactional expertise, resulting in
improved research results for both humanists and nanoscientists.
CONTRIBUTORS:
Brad Allenby, Donna T. Chen, Harry Collins, Robert Evans, Erik Fisher, Peter Galison, Michael E. Gorman,
Lynn Isabella, Lekelia D. Jenkins, Mary Ann Leeper, Roop L. Mahajan, Matthew M. Mehalik, Ann E. Mills,
Bolko von Oetinger, Elizabeth Powell, Mary V. Rorty, Jeff Shrager, Jim Spohrer, Patricia H. Werhane
MICHAEL E. GORMAN is Professor in the Department of Science, Technology, and Society at the
University of Virginia. He is the author of Simulating Science and Transforming Nature.
TRAD I N G ZON E S
GORMAN,
Cover image, bottom circle: Loggerhead turtle escaping a net equipped with a turtle excluder device (TED). NOAA.
editor
AND INTERACTIONAL EXPERTISE
Inside Technology series
THE MIT PRESS

CREATING NEW KINDS OF COLLABORATION
Massachusetts Institute of Technology
Cambridge, Massachusetts 02142
http://mitpress.mit.edu
EDITED BY MICHAEL E. GORMAN
978-0-262-51483-5
Trading Zones and Interactional Expertise
Inside Technology
edited by Wiebe E. Bijker, W. Bernard Carlson, and Trevor Pinch
Janet Abbate, Inventing the Internet
Atsushi Akera, Calculating a Natural World: Scientists, Engineers and Computers during the Rise of
U.S. Cold War Research
Charles Bazerman, The Languages of Edison’s Light
Marc Berg, Rationalizing Medical Work: Decision-Support Techniques and Medical Practices
Wiebe E. Bijker, Of Bicycles, Bakelites, and Bulbs: Toward a Theory of Sociotechnical Change
Wiebe E. Bijker, Roland Bal, and Ruud Hendricks, The Paradox of Scientific Authority: The Role of
Scientific Advice in Democracies
Wiebe E. Bijker and John Law, editors, Shaping Technology/Building Society: Studies in Sociotechnical
Change
Karin Bijsterveld, Mechanical Sound: Technology, Culture, and Public Problems of Noise in the Twen-
tieth Century
Stuart S. Blume, Insight and Industry: On the Dynamics of Technological Change in Medicine
Pablo J. Boczkowski, Digitizing the News: Innovation in Online Newspapers
Geoffrey C. Bowker, Memory Practices in the Sciences
Geoffrey C. Bowker, Science on the Run: Information Management and Industrial Geophysics at
Schlumberger, 1920–1940
Geoffrey C. Bowker and Susan Leigh Star, Sorting Things Out: Classification and Its Consequences
Louis L. Bucciarelli, Designing Engineers
Michel Callon, Pierre Lascoumes, and Yannick Barthe, Acting in an Uncertain World: An Essay on
Technical Democracy
H. M. Collins, Artificial Experts: Social Knowledge and Intelligent Machines
Park Doing, Velvet Revolution at the Synchrotron: Biology, Physics, and Change in Science
Paul N. Edwards, The Closed World: Computers and the Politics of Discourse in Cold War America
Andrew Feenberg, Between Reason and Experience: Essays in Technology and Modernity
Michael E. Gorman, editor, Trading Zones and Interactional Expertise: Creating New Kinds of
Collaboration
Herbert Gottweis, Governing Molecules: The Discursive Politics of Genetic Engineering in Europe and
the United States
Joshua M. Greenberg, From Betamax to Blockbuster: Video Stores and the Invention of Movies on Video
Kristen Haring, Ham Radio’s Technical Culture
Gabrielle Hecht, Entangled Geographies: Empire and Technopolitics in the Global Cold War
Gabrielle Hecht, The Radiance of France: Nuclear Power and National Identity after World War II
Gabrielle Hecht, The Radiance of France: Nuclear Power and National Identity after World War II,
New Edition
Kathryn Henderson, On Line and on Paper: Visual Representations, Visual Culture, and Computer
Graphics in Design Engineering
Christopher R. Henke, Cultivating Science, Harvesting Power: Science and Industrial Agriculture in
California
Christine Hine, Systematics as Cyberscience: Computers, Change, and Continuity in Science
Anique Hommels, Unbuilding Cities: Obduracy in Urban Sociotechnical Change
Deborah G. Johnson and Jameson W. Wetmore, editors, Technology and Society: Building Our
Sociotechnical Future
David Kaiser, editor, Pedagogy and the Practice of Science: Historical and Contemporary Perspectives
Peter Keating and Alberto Cambrosio, Biomedical Platforms: Reproducing the Normal and the Patho-
logical in Late-Twentieth-Century Medicine
Eda Kranakis, Constructing a Bridge: An Exploration of Engineering Culture, Design, and Research in
Nineteenth-Century France and America
Christophe Lécuyer, Making Silicon Valley: Innovation and the Growth of High Tech, 1930–1970
Pamela E. Mack, Viewing the Earth: The Social Construction of the Landsat Satellite System
Donald MacKenzie, An Engine, Not a Camera: How Financial Models Shape Markets
Donald MacKenzie, Inventing Accuracy: A Historical Sociology of Nuclear Missile Guidance
Donald MacKenzie, Knowing Machines: Essays on Technical Change
Donald MacKenzie, Mechanizing Proof: Computing, Risk, and Trust
Maggie Mort, Building the Trident Network: A Study of the Enrollment of People, Knowledge, and
Machines
Peter D. Norton, Fighting Traffic: The Dawn of the Motor Age in the American City
Helga Nowotny, Insatiable Curiosity: Innovation in a Fragile Future
Ruth Oldenziel and Karin Zachmann, editors, Cold War Kitchen: Americanization, Technology, and
European Users
Nelly Oudshoorn and Trevor Pinch, editors, How Users Matter: The Co-Construction of Users and
Technology
Shobita Parthasarathy, Building Genetic Medicine: Breast Cancer, Technology, and the Comparative
Politics of Health Care
Trevor Pinch and Richard Swedberg, editors, Living in a Material World: Economic Sociology Meets
Science and Technology Studies
Paul Rosen, Framing Production: Technology, Culture, and Change in the British Bicycle Industry
Richard Rottenburg, Far-Fetched Facts: A Parable of Development Aid
Susanne K. Schmidt and Raymund Werle, Coordinating Technology: Studies in the International
Standardization of Telecommunications
Wesley Shrum, Joel Genuth, and Ivan Chompalov, Structures of Scientific Collaboration
Charis Thompson, Making Parents: The Ontological Choreography of Reproductive Technology
Dominique Vinck, editor, Everyday Engineering: An Ethnography of Design and Innovation

Trading Zones and Interactional Expertise
Creating New Kinds of Collaboration
edited by Michael E. Gorman
The MIT Press

Cambridge, Massachusetts
London, England
© 2010 Massachusetts Institute of Technology
All rights reserved. No part of this book may be reproduced in any form by any electronic or
mechanical means (including photocopying, recording, or information storage and retrieval)
without permission in writing from the publisher.
For information about special quantity discounts, please email special_sales@mitpress.mit.edu
This book was set in Stone Sans and Stone Serif by Toppan Best-set Premedia Limited. Printed
and bound in the United States of America.
Library of Congress Cataloging-in-Publication Data

Trading zones and interactional expertise : creating new kinds of collaboration / edited by
Michael E. Gorman.
p. cm. — (Inside technology)
Includes bibliographical references and index.
ISBN 978-0-262-01472-4 (hardcover : alk. paper) — ISBN 978-0-262-51483-5 (pbk. : alk. paper)
1. Interdisciplinary research. 2. Scientists—Professional relationships. 3. Technology transfer.
4. Intellectual cooperation. 5. Clearinghouses. I. Gorman, Michael E., 1952–
Q180.55.I48T725 2011
507.2—dc22
2010008710
10 9 8 7 6 5 4 3 2 1
Contents
1 Introduction: Trading Zones, Interactional Expertise, and Collaboration 1

Michael E. Gorman
I Theory 5
2 Trading Zones and Interactional Expertise 7

Harry Collins, Robert Evans, and Michael E. Gorman
3 Trading with the Enemy 25

Peter Galison
4 Interactional Expertise and the Imitation Game 53

Robert Evans and Harry Collins
II Applying Trading Zones and Interactional Expertise to Domains of Practice 71
5 Service Science: A New Expertise for Managing Sociotechnical Systems 75

Michael E. Gorman and Jim Spohrer
6 From Wizards to Trading Zones: Crossing the Chasm of Computers in

Scientific Collaboration 107
Jeff Shrager
7 Authenticity, Earth Systems Engineering and Management, and the

Limits of Trading Zones in the Era of the Anthropogenic Earth 125
Brad Allenby
8 The Evolution of a Trading Zone: A Case Study of the Turtle Excluder Device 157
Lekelia D. Jenkins
9 A Network States Approach for Mapping System Changes 181

Matthew M. Mehalik
viii Contents
10 Embedding the Humanities in Engineering: Art, Dialogue, and a

Laboratory 209
Erik Fisher and Roop L. Mahajan
11 Can Trading Zones and Interactional Expertise Benefit Business Strategy? 231
Bolko von Oetinger
III Ethics and Trading Zones 243
12 Using Trading Zones to Prevent Normalized Deviance in Organizations 245

Michael E. Gorman and Patricia H. Werhane
13 Viewing Trading Zones Developed to Advance Health as Complex

Adaptive Systems 265
Ann E. Mills, Mary V. Rorty, Lynn Isabella, and Donna T. Chen
14 Creating Trading Zones across Continents and Economies: The Female

Health Company 281
Mary Ann Leeper, Elizabeth Powell, and Patricia H. Werhane
15 Conclusion: Future Research on Trading Zones and Interactional Expertise 289

Michael E. Gorman
Index 297
1 Introduction: Trading Zones, Interactional Expertise, and
Collaboration
Michael E. Gorman
The problems and opportunities facing our civilization do not come neatly sorted by
disciplines. This book outlines a framework for fostering collaborations among exist-
ing expertise communities that have radically different views and practices—what
Kuhn called different paradigms, or exemplars (Kuhn 1962). However, the 2006–2011
NSF Strategic Plan (http://www.nsf.gov/pubs/2006/nsf0648/nsf0648.jsp) emphasizes
the importance of collaborating across paradigmatic barriers:
Discovery increasingly requires the expertise of individuals with different perspectives—from
different disciplines and often from different nations—working together to accommodate the
extraordinary complexity of today’s science and engineering challenges. The convergence of
disciplines and the cross-fertilization that characterizes contemporary science and engineering
have made collaboration a centerpiece of the science and engineering enterprise.
Recent National Science Foundation initiatives like nanotechnology emphasize col-

laboration across not only scientific and engineering disciplines, but also the social
sciences and humanities (Roco and Bainbridge 2001, 2002). As Davis Baird articulated
in his testimony before the Senate Committee on Commerce, Science, and Transporta-
tion (May 1, 2003), “Ethicists need to go into the lab to understand what’s possible.
Scientists and engineers need to engage with humanists to start thinking about this
aspect of their work. Only thus, working together in dialog, will we make genuine
progress on the societal and ethical issues that nanotechnology poses.”
The convergence of nanotechnology, biotechnology, information technology, and
cognitive science will produce revolutionary changes in our civilization (Rejeski 2004).
For example, this convergence could change our species in dramatic ways. In the
developed world, minds are already distributed across a wide range of information
devices like laptops and Web servers and BlackBerries; these same technologies make
it possible to create new global cybercommunities (Monge and Contractor 2003).
What would happen if these technologies interfaced directly with a genetically modi-
fied nervous system? Could the convergence revolution lead to greatly extended
2 Michael E. Gorman
capabilities for the wealthy, to the point where they become almost a different species
from the poorest of the poor? Managing this revolution will require collaboration
among ethicists, social scientists, scientists, engineers, policy makers, and stakeholders
around the world (Gorman 2008).
This volume will introduce a new theoretical framework for understanding and
promoting this kind of collaboration across disciplines and cultures. The framework
emerges from cutting-edge work in science, technology, and society (STS), history, and
the psychology of science. The genesis of the volume and the framework was a 2006
workshop on Trading Zones, Interactional Expertise, and Interdisciplinary Collabora-
tion held at Arizona State University’s Decision Theater, a facility designed to encour-
age collaboration among different stakeholders. The workshop was supported by
the National Science Foundation (SES-0526096), the Boston Consulting Group, and
the Consortium for Science Policy and Outcomes (http://repo-nt.sts.virginia.edu/
Tradzoneworkshop/).1 Over four days, participants presented their own research on
and practical experience of the core concepts in the framework. In addition to STS,
areas of expertise included cognitive science, ethics, business, nanotechnology, com-
puter science, policy, and jazz.
My premise in organizing the workshop was that it made sense to try to connect
two new STS ideas that had implications for trading zones. Peter Galison, in his Image
and Logic (1997), had described how trading zones helped him explain how apparently
incommensurable research communities could work together to develop radar and
particle detectors. Then, Harry Collins and Robert Evans (2002) had emphasized that
STS had to deal with the content of expertise, and had proposed a significant new
category: interactional expertise. Based on these ideas, I had written a commentary in
which I suggested how trading zones and interactional expertise could be combined
into a framework for multidisciplinary collaboration (Gorman 2002).
Participants working in a trading zone to develop a new system would have differ-
ent views of that system, based on their interests and areas of expertise. Therefore a
technology in the making, like improved radar or a Mars Rover, might serve as a
boundary object. This concept was introduced to explain coordination of activities
across a diverse set of stakeholders involved in the development of Berkeley’s Museum
of Vertebrate Zoology from 1907 to 1939 (Star and Griesemer 1989). These stakehold-
ers formed what we would now call a trading zone. Annie Montague Alexander, the
museum’s primary backer and director, wanted as complete a collection as possible of
California species, while the zoologist Joseph Grinnell wanted a much more complete
picture of the ecosystem of California and the University of California. Grinnell’s and
Alexander’s goals could be aligned only if specimens were collected with sufficient
Introduction 3
information to determine their place in the ecosystem. Moreover, the collecting was
done primarily by amateurs. One means of coordination was a standardized method
for collecting and labeling specimens, developed by Grinnell, which served as a lingua
franca or creole among the collectors and scientists. The items collected are boundary
objects, in that they are represented differently by participants in the zone. As Star
and Griesemer explain, “In conducting collective work, people coming together from
different social worlds frequently have the experience of addressing an object that has
a different meaning for each of them. Each social world has partial jurisdiction over
the resources represented by that object, and mismatches caused by the overlap
become problems for negotiation” (Star and Griesemer 1989, 412). The standardized
protocol for collecting and recording information allowed certain critical features of
the boundary objects to be captured and translated. The participants in this trading
zone also had to use California itself as a kind of boundary object (see the chapter by
Collins, Gorman, and Evans for details).
Our workshop was an attempt to reflexively create a trading zone around the
framework. The workshop might have proved me wrong: the end result might have
been agreement that the core ideas should not in fact be combined. Instead, vigorous
argument established that a framework could and should be built around these con-
cepts, though the chapters illustrate that the framework is still a boundary object—
participants vary in the degree to which they are aligned around a common
representation. At this stage, this diversity is good, and I hope it will drive improve-
ments in the framework—which is intended to facilitate partial coordination among
a variety of research activities aimed at understanding and facilitating collaboration—
without turning the framework into a “strong program.”
This volume does not attempt to capture the richness of the back-and-forth of
the workshop. (The talks and discussions are available online; send an e-mail to
meg3c@virginia.edu for permission to access them.) By the end, participants were
tracing models on whiteboards. All of the authors in the volume participated in the
workshop, and they incorporated much of this discussion in different ways into their
chapters.
There were other participants who could not find the time to contribute to the
volume. Their ideas have influenced the work reported here in ways that cannot
simply be attributed to individuals (a complete list can be found at http://repo-nt.sts
.virginia.edu/Tradzoneworkshop/guestlist.htm). I am grateful for their participation.
The volume is divided into three sections: theory, applications, and ethical implica-
tions. I provide a brief introduction to each part and a conclusion in which I discuss
implications for future research.
4 Michael E. Gorman
Notes
1. The organizations supporting the workshop are, of course, not responsible for the opinions
expressed by the editor and the authors in this volume.
References
Collins, H., and R. Evans. 2002. The Third Wave of Science Studies: Studies of Expertise and
Experience. Social Studies of Science 32 (2):235–296.
Galison, P. 1997. Image and Logic: A Material Culture of Microphysics. Chicago: University of
Chicago Press.
Gorman, M. E. 2002. Levels of Expertise and Trading Zones: A Framework for Multidisciplinary
Collaboration. Social Studies of Science 32 (5–6):933–942.
Gorman, M. E. 2008. Scientific and Technological Expertise. Journal of Psychology of Science and
Technology 1 (1):23–31.
Kuhn, T. S. 1962. The Structure of ScientificRrevolutions. Chicago: University of Chicago Press.
Monge, P. R., and N. S. Contractor. 2003. Theories of Communication Networks. New York: Oxford
University Press.
Rejeski, D. 2004. The Next Small Thing. Environmental Forum (Washington, D.C.) 21 (2):42–49.
Roco, M. C., and W. S. Bainbridge. 2001. Societal Implications of Nanoscience and Nanotechnology.
Dordrecht: Kluwer Academic Publishers.
Roco, M. C., and W. S. Bainbridge. 2002. Converging Technologies for Improving Human Performance:
Nanotechnology, Biotechnology, Information Technology and Cognitive Science. Dordrecht: Kluwer
Academic Publishers.
Star, S. L., and J. R. Griesemer. 1989. Institutional Ecology, “Translations” and Boundary objects:
Amateurs and Professionals in Berkeley’s Museum of Vertebrate Zoology, 1907–39. Social Studies
of Science 19 (3):387–420.
I Theory
The first chapter that follows is a reprint of an article by Harry Collins, Robert Evans,
and Michael E. Gorman that was written after a 2006 workshop on Trading Zones,
Interactional Expertise, and Interdisciplinary Collaboration. It presents a synthesis of
ideas about trading zones and interactional expertise that emerged, in part, from their
thinking during the workshop and collaboration afterward. This chapter serves as an
introduction to the core concepts of this book and uses them to create a taxonomy
that shows how a trading zone between expertise communities can gradually morph
into a new area of expertise—such as biomedical engineering, which is obviously a
hybrid of biology and medicine. All the participants in the volume have read this
article, and the rest of the chapters play off this theme in different ways, reflecting
the interests, expertise, and research of the authors. The whole volume, therefore,
presents a snapshot of an emerging trading zone.
Peter Galison provides an intellectual narrative of the discoveries and frustrations
that led him to emphasize trading zones. Finding evidence for the locality of practice
in science, he studied the anthropology literature to see how these scientific subcul-
tures, each with its own language, could connect. He discovered that they form trading
zones in which participants create first a shared jargon, then a pidgin, and finally a
creole. Galison points out that scientific disciplines, like languages, emerge out of this
process of hybridization. Galison spends much of his chapter posing objections to the
idea of trading zones and answering them. He even provocatively extends the concept
of a creole to diagrams. He thinks of interactional expertise as similar to the kind of
“out-talk” carried on by scientists and engineers as they work together in trading
zones, discussing practical research problems in ways that experts from another field
can understand.
Robert Evans and Harry Collins discuss recent empirical studies of interactional
expertise that go beyond the workshop. Interactional expertise involves “walking
the talk” of an expert community, which means using the language of the community
6 Theory
in ways that are indistinguishable from the language usage of an expert practitioner.
Collins and Evans conducted a kind of Turing test for interactional expertise, in which
an audience of experts tries to distinguish between an interactional expert and an
actual expert, based on their answers to the same questions. This chapter describes
how other researchers can conduct such tests. Evans and Collins conclude with a
reminder that acquiring interactional expertise takes substantial time and effort; orga-
nizations encouraging and funding collaborations should keep that in mind.
2 Trading Zones and Interactional Expertise
Harry Collins, Robert Evans, and Michael E. Gorman
Trading Zones as the Locus of Incommensurability
Peter Galison introduced the term “trading zone” to the social studies of science.1 His
purpose was to resolve the problem of incommensurability between Kuhnian para-
digms: How do scientists communicate if paradigms are incommensurable?2 Galison’s
approach has two legs. The first leg denies that scientific paradigms are as monolithic
as Kuhn says. The second leg uses the metaphor of the trading zone to explain how
communication is managed where there is a degree of incommensurability. Here we
concentrate on the second leg.
We concentrate on the second leg because the first leg diverts attention from the
interesting philosophical/sociological questions; if paradigms are not monolithic then,
wherever they overlap, there is no problem to be resolved. Thus, Galison points out
that even if there were theoretical incommensurability between, say, the Newtonian
and Einsteinian worlds, experiment went on much as before and those who built
instruments went on much as before. But if Kuhn is read as applying the Wittgenstein-
ian notion of “form of life” to science, then there is less continuity, even in the
realms of experiment and instruments. The “actors” in those spheres may not experi-
ence the continuity that the analyst sees. For example, even if it seems to the analyst
that the same material objects are being built by instrument makers before and after
the revolution, the instrument makers themselves might see them as having different
meanings before and after the revolution, rather as a cowrie shell might be an orna-
ment for one tribe, a unit of currency for another tribe, and your grandmother’s soul
for another tribe. To think of the cowrie shell as the “same thing” in each of these
cases is to privilege a certain kind of analyst’s viewpoint, something which is encour-
aged by the recent obsession in science and technology studies with the material. The
interesting thing is that sometimes different groups manage to trade with material
objects that may be “the same” from our point of view but are not the same from
8 Harry Collins, Robert Evans, and Michael E. Gorman
theirs. Galison makes the point, remarking: “Two groups can agree on rules of exchange
even if they ascribe utterly different significance to the objects being exchanged; they
may even disagree on the meaning of the exchange process itself. Nonetheless, the
trading partners can hammer out a local coordination, despite vast global differences”
(Galison 1997, 783). It is how this is done that is the interesting problem for analysis;
some analysts take apparent material continuity to be the solution to the problem
rather than the problem to be solved. We make no attempt to solve the problem in
this paper, but it should not be forgotten when we talk, below, of “boundary objects.”3
Not all trade is conducted in trading zones—at least, not according to our defini-
tion. We define “trading zones” as locations in which communities with a deep
problem of communication manage to communicate. If there is no problem of com-
munication, there is simply “trade,” not a “trading zone.” Here, however, we consider
only those cases where there are difficulties of communication and ask how they are
overcome. That is the problem of trading zones as we see it. To repeat, if we do not
start with a problem of communication, we do not have the problem of trading zones,
we simply have “trade.”
Interlanguage Trading Zones
To resolve the problem of trading zones as defined here, Galison (1997) looks to real
economic trade in food and other goods between culturally disparate communities.
He claims the problem is solved by the development of “in-between” vocabularies
through which communication can be accomplished. The simplest of these “interlan-
guages” is a “jargon,” more complex is a “pidgin,” while a “creole” is a new language
in itself. Galison applies this metaphor to science. Using it he describes the develop-
ment of technologies such as radar and high-energy-physics particle detectors which
involve/d communication between physicists and engineers, whom he treats as cultur-
ally dissimilar groups. He also describes the growth of new sciences, such as biochem-
istry, which arose out of chemistry and biology. In this case the result is a new
expertise in biochemistry which involves a full-blown creole which can be taught as
a freestanding language/culture to new generations of students.
We can call the resolution of communications via jargons, pidgins, and creoles an
“interlanguage trading zone.” Unfortunately, this model of the trading zone has often
been taken to be the only meaning. But there are other ways to resolve problems of
communication, some of which, we suspect, are much more ubiquitous than the
interlanguage model. We need to find different terms for the different kinds of resolu-
tion and a way of understanding their similarities and differences.
Trading Zones and Interactional Expertise 9
A General Model of Trading Zones
A more general model of trading zones can be developed by considering two dimen-
sions along which trading zones can vary. One dimension is the extent to which power
is used to enforce trade—this is the collaboration-coercion axis. The other dimension
is the extent to which trade leads to a homogeneous new culture—this is the homo-
geneity-heterogeneity axis. The two axes lead, in the familiar way, to four basic types,
ideal versions of which can be represented on a two-by-two table.4 Figure 2.1 shows
the four basic types of trading zone with one or more examples.
Enforced Trading Zones

As can be seen in figure 2.1, interlanguage trading zones are found in the top left-hand
box because, in their ideal form, they involve mutual agreement to trade rather than
coercion, and they tend toward homogeneity in a merged culture. A full-blown creole,
such as biochemistry, is the ideal-type end point. The opposite quadrant—enforced
trading zones—represents the situation with the maximum degree of coercion and the
minimum level of homogeneity. The clearest example is slavery. Here, in the ideal type,
the bargain is entirely one-sided, and no sharing of culture is attempted or even in
prospect. Imagine slaves being used to propel a Roman galley. Trade is being conducted
in the galley: the Romans get their ship propelled, and in exchange the slaves get some
food, some water, and relief from beatings or death. The trade is maintained by physi-
cal force and involves just enough interchange of meaning to allow the bargain to be
understood. This interchange need hardly be cultural at all, as the “terms of trade” can
be “explained” to the human slave in rather the same way as they would be “explained”
to an animal—behavior being “shaped” by physical punishment and reward.5
Homogeneous Heterogeneous
Interlanguage Fractionated
Biochemistry Interactional Boundary
Collaboration Nanoscience Expertise Object
Interpreters Cowrie shell
Peer review Zoology
Subversive Enforced
Coercion McDonald’s Galley slaves
Relativity Use of AZT to treat AIDS
Figure 2.1
A general model of trading zones.
Different in practice and moral complexion, but in the same location in terms of
the two dimensions, are trading zones where the expertise of an elite group remains
“black-boxed” as far as the other participants are concerned. The access of the non-
elite to understanding of the elite culture is tightly controlled by those in power. At
the same time the elite group will make little or no attempt to gain access to the
expertise of the natives. In this model there is no attempt to reduce the cultural het-
erogeneity, and all the control is from the top down. In terms of Collins and Evans’s
“Three Wave” metaphor, the ideology driving such a situation is that of “Wave One,”
with the science and technology being imposed from the top down on the assumption
that, since science gives access to universal truths, adjustments for local conditions
are otiose and there is no reason to meld the expert culture with the nonexpert or
even for the elite to try to understand the nonexpert view of the world (Collins and
Evans 2002). Such a relationship between scientists and nonscientists was once
almost universal.
One example of such an enforced trading zone is the way central planners came
to dominate agriculture and architecture in some parts of the world. This was to prove
disastrous for agricultural production.6 An alternative but more complex example
would be the use of economic and other incentives to change agricultural or industrial
practices in the developing-world countries. In these cases, those on the receiving end
of the policy are obliged to change their practices, for example by growing different
crops using novel methods, but are not required to adopt the cultural viewpoint of
the dominant group. In each case, the constitution of the trading zone is enforced,
although the mechanisms are institutional, legal, and economic rather than brute
physical repression.
Finally, it should be said that there are circumstances in which an enforced trading
zone is beneficial and even morally desirable. It is hard to think otherwise in the case
of, say, the current approach to AIDS in South Africa, where the government has
pressed the case for a healthy diet as an alternative to drugs and in some townships
it is thought that sex with a virgin is a cure.7 It may also be the case in some emergen-
cies. For example, natural disasters that require outside aid and expertise to alleviate
their effects may be better handled without taking time for cultural negotiations even
though more benefit would be gained in the long term if there was time to develop
a more situated approach.
Subversive Trading Zones

Interlanguages take components from both parties to the trade, but communication
can also be achieved when one party’s language overwhelms that of the other. In such
trading zones, two languages, or forms of life, are on their way to being replaced by
one, which originally belonged to just one of the parties; the other language is on the
way to being subverted. A good example is the spread of American-style fast food,
with all that it implies about changes in the rest of society, which has largely occurred
with the acquiescence of the populations involved. The bottom left-hand box in figure
2.1 contains these “subversive” trading zones.
Subversion is another way in which top-down authority in science can work.
Whereas in the enforced trading zone scientific authority was imposed by institutional
means, in a subversive trading zone it gradually supplants the alternatives until it
becomes the socially appropriate response. If one prefers to think of the Einsteinian
revolution as a discontinuity, then one would say that Einsteinian physics language
has colonized Newtonian physics language. This remains the case in spite of the many
Newtonian-looking sentences that are spoken. Every sentence spoken nowadays that
appears Newtonian is really an Einsteinian sentence in disguise, and the speakers know
this (for example, they will say that their sentences refer to special cases of what is in
reality an Einsteinian universe). The same is also true of the ways in which scientific
explanations for everyday events (for example why the sun rises, where babies come
from) come to replace lay and folk theories in most Western societies. We refer to this
kind of process as “cultural subversion.”
In addition to the establishment of cultural hegemony, subversive trading zones
can also be created through technological or physical means. For example, for most
PC users Microsoft’s Windows software is technologically coercive—it is the operating
system that is most readily available. Over time it has become the default option, and
most users use it because they have never thought of using anything else. In this sense,
most PC users are similar to the users of video recorders described by Donald Norman
(1993) who are forced, by the structure and interface of the VCR, to enact the design-
er’s mental model of how the technology should be used. Other possibilities are
written out of this script and are thus never realized in practice. On the other hand,
Apple enthusiasts, or Linux enthusiasts, generally hate Windows and would prefer to
keep the alternative approaches alive. Nevertheless, they may find they have no choice
but to use some of the Windows applications, or at least ensure that their own appli-
cations are compatible with them. Given that they would prefer not to use Windows,
their “trading” with Windows is better described as enforced and belongs in the
bottom right-hand box.
Fractionated Trading Zones

Interlanguage trading zones operate by developing new cultural tools, subversive
trading zones operate by imposing one culture on another, while enforced trading
zones operate with almost no cultural interchange. The final type of trading zone,
which occupies the top right-hand area of the table, involves fractions of cultures as
the medium of interchange. There are two kinds of fractionated trading zones: bound-
ary object trading zones, which are mediated by material culture largely in the absence
of linguistic interchange, and interactional expertise trading zones, which are medi-
ated by language largely in the absence of the material.
It is, of course, not so easy to divide forms of life into their material and linguistic
components. We have already argued that, as with the cowrie shell, what may appear
materially “the same” to an analyst may appear different to the natives occupying
different forms of life, so treating each “fraction” in isolation presents conceptual dif-
ficulties. As Galison (1997) explains, ideas such as “boundary object” work because,
though the object means different things to the different parties, this does not vitiate
their separate projects. One well-known example of a boundary object trading zone
is that described by Star and Griesemer, who studied how a diverse group of scientists,
trappers, amateur collectors, and university administrators successfully collaborated
in providing and cataloguing specimens for the Museum of Vertebrate Zoology at the
University of California, Berkeley. According to Star and Griesemer, “Boundary objects
are objects that are both plastic enough to adapt to local needs and constraints of the
several parties employing them, yet robust enough to maintain a common identity
across sites” (Star and Griesemer 1989, 393). One of the boundary objects in their
study was the state of California:
The maps of California created by the amateur collectors and the conservationists resembled
traditional roadmaps familiar to us all, and emphasized campsites, trails and places to collect.
The maps created by the professional biologists, however, shared the same outline of the state
(with the same geo-political boundaries), but were filled in with a highly abstract, ecologically-
based series of shaded areas representing “life zones,” an ecological concept. (ibid., 411)
Both groups could therefore agree that specimens should be collected from California,
and location information could be translated between the two kinds of maps if the
right data were provided. Standardized procedures for recording information facili-
tated this translation. In effect, the boundary objects plus the standardized procedures
performed the same function as a creole.
The linguistic complement of the boundary object is interactional expertise. This
involves internalization of the tacit components of a strange language. While acquisi-
tion of interactional expertise does not provide a full grasp of the strange form of
life—it provides no access to the other parties’ material culture except insofar as that
material culture is represented in discourse—it is surprising how much can be done,
is done, and, indeed, must be done, with the language fraction alone. Interactional
expertise is the product of a successful linguistic socialization. Although expressed as
language alone, it cannot be too heavily stressed, interactional expertise is tacit-

knowledge-laden and context-specific. It is not the same as the kind of formal repre-
sentation of the world beloved of artificial intelligence enthusiasts and the like.
Language as interactional expertise is just as resistant to representation in computer
programs as any other kind of tacit and situated knowledge.
One example of what seems to be a fractionated trading zone based on interac-
tional expertise is water management in Arizona. To make a decision about appro-
priate technology, social scientists were recruited to develop a set of metrics that
would allow Bureau of Reclamation engineers, Yavapai Indians, and other stakehold-
ers to communicate about the value, use, and need of water resources (Espeland
1998). We could say that the social scientists were engaged in two simultaneous
projects: trying to become interactional experts in the worlds of the various nego-
tiating parties and trying to build an interlanguage—in this case a common language
of measurement—that both parties would share. The social scientists failed to build
the interlanguage; both the Bureau engineers and the Yavapai felt that the metrics
developed by the social scientists failed to capture important parts of their separate
beliefs and values. The interactional expertise part of the enterprise seems to have
worked, however. A compromise was reached that avoided the building of a new
dam, preserved Yavapai land, and still permitted water management (ibid.). Cultures
that remained incommensurable from a cultural standpoint seem to have found a
way to trade via the intervention of social scientists who acquired interactional
expertise in both worlds.
Another example of the use of interactional expertise to create a trading zone is
that of the San Francisco AIDS activists described by Steven Epstein (1996). The AIDS
activists mastered the language of medical research even though they were not medical
researchers or practitioners themselves. Initially, groups of activists demanded changes
in the protocols and researchers made some concessions, although this relationship
was more adversarial than cooperative. As activists grew in sophistication and attracted
allies among statisticians, however, the relationship became more collaborative,
although both groups retained their separate identities and cultures.
A third example of a fractionated trading zone is provided by the development of
an environmentally intelligent fabric by a team that included an architect, a green
chemist, a fashion designer, and the manager of a textile mill (Gorman and Mehalik
2002). The fashion designer recruited a “high practitioner,” an architect who pro-
nounced that the fabric would have to follow the maxim “waste equals food.” The
maxim was based on an analogy with natural systems, claiming that all waste in nature
is food for other organisms. “Waste equals food” became a provocative metaphor for
the growing team. The green chemist added a protocol for determining which materi-
als and processes fulfilled this maxim. The mill owner had to create the fabric in
alignment with the green chemist’s protocols. All core members had to achieve inter-
actional expertise in each other’s domains in order to produce a successful commercial
product. Thus, for example, the green chemist and the architect had to understand
enough of the manufacturing processes to be sure they followed the protocol, while
the mill owner had to understand the guiding principles and protocol well enough to
be able to embody them in the design. Nevertheless, the parties did not become con-
tributors to each other’s fields, nor was any new vocabulary developed—at least, not
in the first instance.
Yet another example is the field of gravitational wave detection. This is a highly
integrated esoteric science, in which everyone is committed to the common goal of
building a gravitational wave detector; there are perhaps 500 gravitational wave
scientists. The field is divided up, however, into many subspecialisms, most scientists
have only interactional expertise in the work of a specialism that is not their own,
and there do not seem to be any new, gravitational-wave-specific, interlanguages.
Could it be that the growth of interlanguages is the unusual case? It may be that,
when examined closely, what appear to be integrated networks of scientists are really
conglomerations of small groups bound together by rich interactional expertises.
In these examples trades were managed even though interlanguages were not the
key. Goods which had different values in each culture were traded, the exchange being
managed either by third parties who had the capacity to talk to both in some approxi-
mation to their language or by members of each group gaining interactional expertise
in others’ worlds. In no case was there full immersion (i.e., practical as well as linguistic
socialization) in the others’ forms of life or the development of new languages
or cultures. The interactional expertise trading zone seems so widespread that it
might be argued that this, rather than the interlanguage model, is the norm for new
interdisciplinary work. Certainly, where interlanguages develop it is likely that
interactional expertise is the first step.8
Trading Locations and the Fractal Approach
One potentially misleading feature of the left-hand areas of figure 2.1 is that once the
examples of trading zones that belong there reach their end point—a full-blown creole
with a degree of institutional stability such as is exhibited by biochemistry, or a total
swamping of one native culture by the other—the problem of communication has
been solved and, according to our usage, the “trading zone” disappears. Thus biochem-
istry, though it grew up as a trading zone, is now just a new homogeneous cultural
location in which trades happen. When they reach their end points, all the examples
in the left-hand areas slip off the table in a westerly direction, as it were.
The question remains of what these groups consist of once they have achieved
the appearance or reality of cultural homogeneity. We have already suggested that a
uniform-looking area such as gravitational wave physics is actually made up of many
small local groups interacting through interactional expertise. There is also a more
“conceptual” point. It is possible to find discontinuity even in as solidaristic a cultural
location as biochemistry by looking closely enough. One can always choose to “zoom
in” on any area of social life and, as the scale increases and ever more detail is
exposed, as with a polished metal surface, what appeared smooth turns out to be
jagged. Social life, one might say, is like a fractal where the structure is reiterated at
every scale and the scale at which one chooses to pitch one’s analysis is a matter of
choice. Thus, within the mature science of biochemistry there will be small and local
regions that still exhibit the characteristics of all the kinds of trading zone represented
in figure 2.1.
This fractal model comes naturally to psychology, though not by that name. Psy-
chologists talk less in terms of cultural similarity and discontinuity and more in terms
of “shared mental models.” For such psychologists the fractal model is already imma-
nent in the way they look at the world, because the sharing may be among greater or
smaller numbers of individuals. Thus the idea that a large group of people may share
a large-scale mental model while not all of them share aspects of its more detailed
structure is accepted without special mention. Whether one talks of fractals or differ-
ent aspects of mental models being shared by more or fewer individuals is a matter
of choice of language. In this case, the different languages do not appear to present a
problem of communication, as it is possible to translate between the two different
discourses without loss. Thus, for example, what a sociologist might see as a low degree
of inclusion within a technological frame, a psychologist might describe as a partially
shared mental model.9
On the right-hand side of the diagram there is no slipping off. On the right-hand
side, trading zones remain trading zones without any need for magnification even
after long periods of evolution because the separate cultures remain separate. In the
case of interactional expertise, one party learns the language of the other while
retaining their own material form of life and distinct contributory expertises. Indeed,
it is precisely the continuing discontinuity between the cultures that enables the
individual with interactional expertise, who thus has a mastery of both languages,
to maintain his or her special role. For example, interpreters can do their job via
interactional expertise, going backward and forward between the two groups only so
long as the two groups want to communicate but are unable to do so.10 In the case of
boundary objects the two forms of life again remain distinct, each imposing their own
meaning on the “common” material object and working with it in their own way. In
the case of slavery, two distinct cultures also remain. In the extreme case of the Roman
galley, the slaves are given no opportunity to learn the culture of their masters, and
the trading zone remains unchanged for as long as the slave masters can exert their
will. In more complex settings the continuation of the two cultures may be harder to
see, but it is nonetheless still present. For example, where slaves perform more complex
tasks they appear to be participating more fully in the dominant culture, so it might
appear that the slaves’ culture is being replaced by that of their masters (i.e., the
trading zone is becoming a subversive one). In practice, however, a closer inspection
often reveals that the original culture continues to survive, albeit in less visible ways
and places, so that the trading zone continues to be based on two distinct but differ-
ent cultures even if, on the surface, it appears as just the one.11
Many of the most interesting instances of interactional expertise—such as peer
review and the management of scientific projects—can be seen as features of a trading
zone only when the magnification is high. Viewed from a distance—from high up the
fractal—they look like elements within a coherent scientific culture rather than mecha-
nisms that allow disparate cultures to communicate.
The Evolution of Trading Zones
If trading zones are dynamic entities, then it should be possible to use the typology
set out above to describe the different states in which a trading zone might exist over
time. While it is possible for new domains to be so innovative that they do not build
on any existing cultures, this situation is probably rare. One example may be the
ARPANET, the precursor to the Internet. Here there was no problem of incommensu-
rability between the partners because the project was so new that the culture had to
be invented jointly from a blank slate. Democratic organization and trust appears to
have been sufficient for coordination.12
By considering the twin dimensions of homogeneity-heterogeneity and collabora-
tion-coercion, we can analyze the different trajectories and characteristics of trading
zones and the way these change as trading zones develop. This possibility is best
visualized on a space that admits of in-between positions as well as ideal cases. Figure
2.2 represents such a space, with degrees of coerciveness and homogeneity now rep-
resented on the continuous axes of a graph. It shows the possible trajectory of an
INTERLANGUAGE COLLABORATION FRACTIONATED
3 2a 2b
Interlanguage Boundary Interactional
object expertise
TRADE
HOMOGENEITY HETEROGENEITY
1
“Encouraged” to collaborate
4 5
Homogeneity Institutional power
SUBVERSIVE COERCION ENFORCED
Figure 2.2
Evolution of a trading zone.
imaginary research group coming together around the idea of nanotechnology. The
trajectory follows a counterclockwise movement starting in the bottom right-hand
quadrant, each succeeding stage being indicated by a number from 1 to 5. The trajec-
tory is not intended to be prescriptive, it is merely an illustration of the way in which
the categories set out above might be applied in a more dynamic context.
In this example, step 1 arises when a university “encourages” the members of its
chemistry, physics, engineering, and social science departments to collaborate on a
research application to develop “nanoscience.” At this point disciplinary identities are
distinct, so they remain heterogeneous rather than homogeneous, and their coming
together has been enforced, albeit gently, so the trading zone is coerced rather than
collaborative at this stage. At this point, there are two main ways in which the trading
zone might develop. On the one hand, the group may decide to continue to collabo-
rate and develop a research proposal; or they may conclude that there is insufficient
common ground, and the trading zone would then collapse.13
Assuming the group decides that working together is viable and desirable, the
trading zone would become increasingly voluntary and collaborative. In this scenario,
the different departments agree that the nanoscience application is a good idea and,
even though they remain separate disciplines, the initial coercion is replaced by a
shared agreement to work together. The collaboration thus migrates upward and
becomes a fractionated trading zone of one of two types. One outcome—step 2a in
figure 2.2—is that some kind of boundary object (in this case the research application)
is produced.14 Although it is likely, given the continued existence of different depart-
ments and disciplines, that the research application will mean different things to
different people, these differences are not sufficiently important to undermine the
joint project. As work on the application develops and the different parties become
increasingly engaged in the ideas, we might expect to see the trading zone drift further
up the diagram as the intensity of the collaboration increases.
A shared research proposal is not the only outcome, however. It may be that,
despite the initial optimism, no application is produced or, even if it is, that it is not
funded. In these circumstances the initial trading zone might disintegrate, as the
object that united them has now disappeared. On the other hand, it may be that, as
a result of the interactions and the sharing of ideas and discourse, some members of
the trading zone become sufficiently interested in the others’ work to want to under-
stand more about it. If they pursue this ambition, it is possible that they will develop
interactional expertise in one of the other disciplines—step 2b in the figure—and that
a fractionated trading zone based around this shared expertise will emerge.
If, however, we assume that the trading zone develops into the boundary object
type, united around a successful research application, then the initial collaboration
will continue. At this point, as the nanoscience work begins in earnest, the interac-
tions between the different departments and disciplines will intensify still further.
Now they may be co-located in the same building, working on joint projects. Research-
ers will thus have to communicate and coordinate their actions with a new and diverse
set of colleagues. They might start with interactional expertise as their main commu-
nicative resource and then begin to invent jargon terms, which in turn give rise to a
pidgin and, perhaps, a creole. As this happens, so the cultural differences between the
different disciplinary participants will be reduced and the amount that they share will
increase. As cultures become more homogeneous, so the trading zone begins to move
to the left and the fractionated trading zone becomes an interlanguage trading zone—
step 3 in the figure. Over an even longer time period it may even be the case that, as
with biochemistry, a distinct new discipline emerges. In figure 2.2, this would be
represented as a further drift to the left to the point where there is so much in common
that the trading zone concept does not apply. There is simply a new science in which
trade takes place in the ordinary way.
The creation of a new strong science sets the scene for other changes too. One
possibility is that, as the new science develops and becomes more popular, then, just
as the language of Newtonian physics was eventually subverted by the paradigm shift
to Einsteinian relativity, so our current understanding of the boundaries between
information science, genetic science, and the natural sciences as a whole will be sub-
verted by the new language. This possibility is realized when departments are set up,
textbooks are produced, and what was once a radical and innovative experiment
becomes a normal science research program. All this takes place over the western
horizon of the figure. New recruits to the discipline will now find the cultural hege-
mony so strong that they have no choice but to abandon any ideas they bring with
them and accept the dominant culture. Anything strange sucked into its domain will
be subverted by the dominant culture of the new science. We represent this with the
label “hegemony” in the bottom left-hand quadrant.
The creation of a strong science also changes the way in which the science interacts
with other groups and becomes part of new trading zones. For example, as the new
nanoscience becomes more institutionalized within the broader network of scientific
disciplines, so the elite of the new science will need to recruit new scientists and police
the boundaries of the new field in order to control who should count as an expert in
that domain. This institutionalization, where it happens, is shown as step 5 in the
figure and returns us to the bottom right-hand corner, where the authority of expertise
allows the discipline to make knowledge claims that can now be enforced over other
knowledge claims, for good or for bad, in the manner described above.
We must stress again that this is just one possible model and it is certainly not a
prescriptive one. It may be that the trading zone simply ceases to function and disin-
tegrates somewhere along the way. For example, a team, having reached step 2, might
disband when the problem it was created to solve has been dealt with, or it might be
that the difficulties of interdisciplinary collaboration become too great as the detailed
work begins to be done. Similarly, either of the fractionated trading zone states may
prove to be a stable equilibrium so that the development of an interlanguage trading
zone is neither necessary nor desirable.
The anticlockwise direction sketched above is also just one possibility. It may be
that the power dynamic of the trading zone changes such that one party becomes
increasingly able to enforce their views on the others. This would be represented by
a drift downward (as coercion becomes more relevant) and possibly to the left (as the
differences are erased). In the case of the nanoscience research project, this outcome
might represent the capture of the research agenda by one of the subdisciplines and
an increasing emphasis on a single perspective, to the point where nanoscience
becomes a subfield of chemistry or another existing disciplinary culture. In other

words, there is no particular “right” way of moving through or around the figure, but
the differences between different kinds of trading zones can be made clearer by think-
ing about how they can be represented in the two-dimensional space it defines.
Funding Trading Zones
If figure 2.2 were drawn to scale, then (at least for science and technology) the top
half would be the largest and most important, emphasizing the extent to which inter-
actional expertise and cross-disciplinary working are central to most modern technical
collaborations. By considering how these different groups were brought together in
the first place (e.g., by choice or compulsion) and the extent to which they remain
distinct, this model provides a framework for understanding the problems that might
arise. For example, in the case of enforced trading zones, motivation and compliance
are likely to be problematic. Given the emphasis on collaboration, this is less likely
to be a problem in the top half of the diagram. Moreover, to the extent that genuinely
novel ideas come from “left field” or from those with a relatively low degree of inclu-
sion, then the top right-hand quadrant seems to be the best location for developing
new interdisciplinary partnerships.
There is also a risk with this activity, however. The fractionated trading zone identi-
fies a kind of research activity that funding bodies need to support but to support
charitably. Given the heterogeneity of the cultures and the difficulty of learning new
languages, the realistic expectation must be that only a small number of such trading
zones will succeed, either as stable fractionated trading zones or as the more institu-
tional homogeneous interlanguage variety. In funding nascent fractionated trading
zones, a good allowance must be made for the sheer time it takes to acquire interac-
tional expertise and/or develop boundary objects that are both malleable and robust
enough to bridge between the different communities. Yet longer time frames must be
used if the aspiration is the emergence of pidgins and creoles needed for new inter-
languages to develop.15
Conclusion
The idea of a trading zone as a place where problems of communication and coordina-
tion are resolved can help us understand a wide range of styles of social and scientific
collaboration and the ways in which they may evolve into one another.
In this chapter, we have argued that the different possibilities can be captured in
a two-by-two matrix based on the axes of coercion and homogeneity. These ideal
types provide exemplars of the different ways in which trading zones can be made to
work. They are:
• Interlanguage trading zone (high collaboration, high homogeneity),
• Subversive trading zone (high coercion, high homogeneity),
• Enforced trading zone (high coercion, high heterogeneity),
• Fractionated trading zone (high collaboration, high heterogeneity).
We have also argued that, using the same axes, it is possible to describe how a trading
zone may develop over time by moving between different states. This last point is
particularly important in the case of interdisciplinary science. Firstly it shows that
there is not just one best way of organizing interdisciplinary collaborations and that,
even within the same collaboration, different relationships will develop at different
times. Secondly, and perhaps even more importantly, thinking about trading zones
as places where cultures meet, languages are learned, and tacit knowledge shared
emphasizes the difficult and time-consuming nature of the work. This is not to say
that such work should not be funded, but it does serve as an important reminder
that such work is different from that of normal science and needs to be managed
and assessed in different ways. In particular, to the extent that fractionated trading
zones depend on the development of interactional expertise, they require a consider-
able work on the part of at least some participants if their potential is to be
realized.
Notes
The overall approach offered here is inspired by Gorman 2002, where a classification of trading
zones based on the idea of interactional expertise is developed. We also draw heavily on a paper
by Ribeiro (2007a), who provides a model of communication with many of the components
discussed here. The writing of the paper was inspired by a workshop on trading zones and inter-
actional expertise organized by Gorman at the University of Arizona, May 21–24, 2006, and
supported by the National Science Foundation (SES-0526096), the Boston Consulting Group, and
the Center for Nanotechnology and Society.
1. The key reference for this is Galison 1997.
2. Kuhn introduces the term paradigm in his classic The Structure of Scientific Revolutions (Kuhn
1996).
3. The standard reference for boundary objects is Star and Griesemer 1989. An alternative way
to explore the problem is to analyze interaction across cultures at the level of human actions
(Collins and Kusch 1998, ch. 4). Collins and Kusch argue that actions cannot be coordinated
across cultures unless the actions are “mimeomorphic,” which is to say that repeating the exter-
nally describable behaviors associated with an action is sufficient to reproduce the effect of the
action irrespective of the intentions. “Polimorphic” actions, in contrast, need to be understood

if their effects are to be reproduced, since context is relevant and mere reproduction of behavior
will fail. Thus, it is the mimeomorphic components of “action trees” associated with the handling
of cowrie shells that allow trade.
4. Gorman and Mehalik earlier proposed three stages in trading zones, the first of which cor-
responds to coercion, the second to our collaboration categories above, and the third to the kind
of shared mental model achieved by the inventors of the ARPANET and by other design and
discovery teams on the frontiers, where boundaries and terminologies are fluid. For more details
of this typology, see Gorman and Mehalik 2002.
5. No doubt more complex tasks would require a great deal more in the way of cultural inter-
change if they were to be successfully carried out by slaves, but we are describing the case in
such a way that fills the bottom right-hand quadrant in the manner of an “ideal type.”
6. For discussion of this and other cases, see Scott 1998.
7. For an examination of an earlier example of the South African government’s sometimes bizarre
approach to the AIDS epidemic in its country, see Weinel 2007.
8. For more on this, see Shrager 2007.
9. The concept of a technological frame comes from Bijker 1995.
10. For more on the work of interpreters, see Ribeiro 2007a, 2007b.
11. The same is true in those cases where slave masters make a deliberate attempt to eliminate
the indigenous culture. Where they succeed, the trading zone becomes a subversive one; where
they fail, it remains an enforced one.
12. For more on the development of ARPANET, see Hughes 1998.
13. Logically there is a third possibility, which is that the university authorities would refuse to
accept the decision that a research proposal could not be developed and would insist that the
collaboration be made to happen. In this case the trading zone would move south on the diagram
as the relative degree of coercion increased.
14. Anecdotally it seems from presentations at the Arizona workshop that one of the typical
stumbling blocks in these kinds of processes is the physical location of the research work. In
most cases it seems that working in a brand-new building is better than working in one that is
already seen as “belonging” to one of the contributing departments.
15. For a discussion of interdisciplinary research in the context of sustainable cities research, see
Evans and Marvin 2006.
References
Bijker, W. E. 1995. Of Bicycles, Bakelites, and Bulbs: Toward a Theory of Sociotechnical Change.
Cambridge, MA: MIT Press.
Collins, Harry, ed. 2007. Case Studies of Expertise and Experience. Special issue of Studies in
History and Philosophy of Science 38 (4).
Collins, Harry, and Robert Evans. 2002. The Third Wave of Science Studies: Studies of Expertise
and Experience. Social Studies of Science 32 (2):235–296.
Collins, Harry, and Martin Kusch. 1998. The Shape of Actions: What Humans and Machines Can
Do. Cambridge, MA: MIT Press.
Epstein, Steven. 1996. Impure Science: AIDS, Activism, and the Politics of Knowledge. Berkeley:
University of California Press.
Espeland, Wendy N. 1998. The Struggle for Water: Politics, Rationality, and Identity in the American
Southwest. Chicago: University of Chicago Press.
Evans, Robert, and Simon Marvin. 2006. Researching the Sustainable City: Three Modes of Inter-
disciplinarity. Environment and Planning A 38 (6):1009–1028.
Galison, Peter. 1997. Image and Logic: A Material Culture of Microphysics. Chicago: University of
Chicago Press.
Gorman, Michael E., and Matthew M. Mehalik. 2002. Turning Good into Gold: A Comparative
Study of Two Environmental Invention Networks. Science, Technology and Human Values 27
(4):499–529.
Hughes, Thomas P. 1998. Rescuing Prometheus: Four Monumental Projects That Changed Our World.
New York: Vintage Books.
Kuhn, Thomas S. 1996. The Structure of Scientific Revolutions. 3rd ed. Chicago: University of
Chicago Press.
Norman, Donald A. 1993. Things That Make Us Smart: Defending Human Attributes in the Age of
the Machine. New York: Addison Wesley.
Ribeiro, Rodrigo. 2007a. The Language Barrier as an Aid to Communication. Social Studies of
Science 37 (4):561–584.
Ribeiro, Rodrigo. 2007b. The Role of Interactional Expertise in Interpreting: The Case of Technol-
ogy Transfer in the Steel Industry. In Collins 2007, 713–721.
Scott, James C. 1998. Seeing Like a State: How Certain Schemes to Improve the Human Condition Have
Failed. New Haven: Yale University Press.
Shrager, Jeff. 2007. The Evolution of BioBike: Community Adaptation of a Biocomputing Plat-
form. In Collins 2007, 642–656.
Star, Susan Leigh, and James R. Griesemer. 1989. Institutional Ecology, “Translations” and
Boundary Objects: Amateurs and Professionals in Berkeley’s Museum of Vertebrate Zoology,
1907–39. Social Studies of Science 19 (3):387–420.
Weinel, Martin. 2007. Primary Source Knowledge and Technical Decision-Making: Mbeki and
the AZT Debate. In Collins 2007, 748–760.
3 Trading with the Enemy
Peter Galison
One way to think through what a concept like the trading zone does is to press objec-
tions against it, for only then do sharpened boundaries pull foreground from back-
ground. Analyzing such confrontations tracks my ideas about these scientific
subcultures and exchange languages. But because it is sometimes useful to start with
the history of a concept, I want to begin there—and then follow the history into more
analytical territory.
What grabbed me most in Marx’s work—and the history of work more generally—
was certainly not the labor theory of value and the interminable battles over its limits.
Instead, what impressed me were the discussions of machines: the descriptions of
looms and labor, the vivid depiction of how bosses drove down the number of cubic
feet of air that weaving girls had in their quarters. Among the historians who were
current when I was starting out, it was the work of the Annales School I liked best:
the history of how medieval land was ploughed (Marc Bloch); how rice fields were
easier to police than the hill towns of Tuscany (Fernand Braudel). I liked seeing how
work worked—how cars were pounded together, mine faces stripped of coal, and
secretarial work narrowed. Studies like those by Harry Braverman (Labor and Monopoly
Capital, 1974) intrigued me; so too did the great historical studies by E. P. Thompson
in The Making of the English Working Class (1966).
It was the actual scientific work that I wanted to get at in writing about the history
of science—and such a history of science seemed impossible to achieve if one ignored
the laboratory. I was utterly transfixed by these experimental spaces; I had spent a
year in a plasma physics lab studying ion waves, and several months in an applied
physics lab trying to figure out how to best spray water to keep a miniature, idealized
house from burning to the ground. I had studied with a truly great experimental
physicist, Robert Pound, and watched, riveted, as he, a true master, plucked electrical
signals out of the noise.
26 Peter Galison
In How Experiments End (1987), I wanted, above all, to capture the weight that
experimental practice had as a distinct form of reasoning—a form of reasoning not
reducible to inspiring theory or checking after the fact. On the contrary, the point
was to show how experiments really did move to a rhythm distinct from that of
theory, that experimentalists’ decision that they’d seen something real (for example)
was not grounded on the same standards and forms of argumentation that satisfied
theorists that they had found a bona fide effect. It was this quasi-autonomy that led
me away from the then overwhelmingly popular Kuhnian picture of mutually incom-
prehensible paradigms. I just didn’t see the experimentalists finding incommensurabil-
ity in their practices before and after theoretical breaks such as the 1905 advent of
special relativity.
During that period—late 1970s and early 1980s—the laboratory and the experiment
were discussed more often in science studies. But as much as I objected to the mar-
ginalization of experiment in favor of theory, I also bridled at the reanimated reductive
form of positivism that dismissed theory and theorists, placing reality in experiment
above all else. Theory, like experiment, had its own culture of demonstration, its own
short-, middle-, and long-term constraint structure that characterized what it meant
to be a theorist. By the time I published How Experiments End, I had a picture of three
intercalated, quasi-autonomous subcultures of theory, experiment, and instrument
making.
So far, so good. But then I got good and stuck. Here was the problem. On the one
hand we had the Kuhnian picture of paradigmatic splits—revolutions—that thor-
oughly and unbridgeably cleaved science onto one side or the other of a great divide.
This view was taken up with increasing frequency even among my allies in the new
and burgeoning field of laboratory studies. On the other hand, I saw the weight given
to experimental culture as pulling in another direction—toward the intercalated
picture of three subcultures that I hoped would better capture the phenomenology of
scientists’ experience—scientists who seemed very rarely to have seen themselves as
forever banished from the far shores created by a putative epistemic split.
In 1988, I reported on where I was with this train of thought about intercalation,
rupture, and continuity in my essay “History, Philosophy, and the Central Metaphor.”
In writing this piece for Science in Context, I found that the intercalated periodization
not only failed to resolve the incommensurability problem, but made it much worse.
For Kuhn had used three criteria to pick out a paradigm-bearing community of scien-
tists. First, the community shared a basic agreement about what there was in the world
and how these things interacted (ontology); second, they held in common a set of
acceptable means for learning about these entities (epistemology); and third, they
shared an understanding of basic physical laws (nomology). Alongside this framework
Trading with the Enemy 27
of knowledge was its sociological support: scientists within a community had shared
routes for circulating knowledge, such as preprints, conferences, and journals. But
if the experimentalists were to be autonomous enough to cut across the theorists’
paradigms—as I saw they often were—it was precisely because they did not share the
full-bore commitments of the theorists to the nature of objects, laws, and ways of
acquiring knowledge; and they certainly had a large body of their own conferences,
preprint exchange networks, and journals.
It followed that what I had on my hands was a picture of both a diachronic incom-
mensurability of the Kuhnian sort and a synchronic incommensurability between
experimentalists and theorists (for example) at any particular time. I had wanted a
picture of layered strength, like a New England stone wall; instead, I got one made of
fragile, delaminating plywood. I was stuck, banging my head against this rock wall.
By 1989, I had written various pieces of a book (Image and Logic, 1997) about the third
subculture (instrument making), and it was ever clearer that I had painted myself into
a corner: the more I argued that the three subcultures had a kind of autonomy from
one another, the more blatant it became that I had no idea how to address the way
that cross-talk might actually link them.
If the incommensurability problem was harder to solve, it was at least increasingly
clear to state. The most important lesson of the previous decade, it seemed to me, had
been the locality of practice. I thought of theory as having its own form of local
practices—for example, I had long been interested in how Hermann Minkowski’s
Goettingen-mathematical way of formulating relativity theory differed from Einstein’s
own. Conversely, I was fascinated with the characteristic ways that experimentalists
handled their forms of argumentation and demonstration in the laboratory. But our
ways of talking about language in the history of science were anything but local.
Instead, they were, through and through, global. W. V. O. Quine and Thomas Kuhn,
Rudolf Carnap and Otto Neurath—even (or especially) in the new (neo-Kuhnian)
sociology of science of the 1980s, we were still talking about global translation. My
specific problem was that I had boundaries this way and that (diachronic and syn-
chronic, among three subcultures) and no good way of showing anything local about
what happened at the various intersections.
Put differently: by the late 1980s, we had an increasingly adequate account of
local practice and were joining it, unworkably, to a global account of language. That
was the difficulty. And the way to address it was to figure out how to talk about
language as practice—local language practices. So I began poking around in the lit-
erature of anthropological linguists, who I hoped would have something to say about
languages in border zones. What I found was perfect for what I wanted: the anthro-
pological linguists had indeed studied such situations. Soon I came across a book,
28 Peter Galison
Tom Dutton’s Police Motu: Iena Sivarai (Police Motu: Its story, 1985), which I cottoned
to immediately. Dutton not only showed the ways that a hybrid pidgin formed, but
tracked it across time, showing how a “natural” exchange language could become
more and more expansive until it could function in full, used in radio programs and
everyday life.
I began thinking about more partial kinds of scientific languages, of the work
needed to produce the first collective and coordinative moves that would join biology
to chemistry, and slowly articulate a language that borrowed from both but was
subservient to neither. Why not think of these hybrid arenas of practice as a form
of language? After all, on what grounds would one dismiss as beyond the pale of
“real” language the highly restricted call-out system of the type Wittgenstein identi-
fies as a language game in the Philosophical Investigations? Are Fortran or C++ “just
metaphors” for a language? And if algebra and geometry count as languages, why
not count algebraic geometry? When a physicist says, “We can say that in the lan-
guage of differential geometry or in the language of quantum field theory,” is this
utterance purely metaphorical? On what grounds would one base such a high-handed
dismissal?
For analytic, functional, and political reasons, the anthropological linguists were
committed to the idea that pidgins and creoles were languages, not inferior or debased
forms of “real” or “natural” languages. Indeed, anthropological linguists were irritated
for a reason. In the days before Chomsky, linguists used to treat pidgins and creoles
as “simple,” not “true” languages; in fact, it used to be a commonplace to rank lan-
guages from complex to simple. (I remember reading many years ago the early modern
mathematician and engineer Simon Stevinus’s confident assertion that Dutch was the
best possible language for science. Finding French authors who judge French to be the
clearest of possible voices for reason is not very hard—for that matter, we need no
commission of inquiry to locate the extensive German literature on the superiority of
German for poetry, music, or philosophy.) Happily, linguistics has given up the
ranking of languages along with the systematic demotion of interlanguages like the
various forms of pidgin English.
Why should science studies plant its flag in a nineteenth-century conception of
language? No contemporary linguist takes Dutch, French, German, or English as
“pure” these days—everyone knows that linguistic hybridity goes all the way back.
Contemporary English, for example, contains a long series of borrowings, intrusions,
and mixtures, inter alia from Anglo-Frisian dialects, Scandinavian German, Old
Norman, Latin, and Greek—and the adoptions and adaptations continue unabated.
Are there periods of relative stability? Of course. So it is in science, too. Classical
physics is not pure in some originary sense. We know perfectly well that it was formed
out of a complicated admixture of mathematical statics, craft practices, natural phi-
losophy, printing technologies, and much else besides. So why look down on chemical
physics or physical chemistry as the impure, lesser form of science or scientific lan-
guage? To pursue this opening allows us to ask many more questions. For example,
what pieces of physics and what pieces of chemistry are brought together? Where do
the calculational procedures originate? Which laboratory procedures are brought into
the combination and which are left behind?
Understanding the locality of interlanguages—and how they might be applied to
science—cracked the final theoretical obstacle to linking practices of theorizing, instru-
ment making, and experimenting. I could now see the way “out-talk” functioned
when experimentalists addressed theorists. I could ask questions sharply: What exactly
is left behind and what put forward when an instrument maker addresses an experi-
mentalist? In other words, what forms of regularization occur in the scientific registers
of jargons, pidgins, and creoles? What did theoretical physicists hold back and what
did they put on the table when they talked to radio engineers when the two groups
tried to build radar during World War II? It took another eight years for me to com-
plete Image and Logic (long story, long book), but by the fall of 1989 the framework
for treating local trade and trade languages—rather than global translation—was in
place.
Over the years, I grappled with many objections as I came to terms with the trading
zone and its associated exchange languages.
Objection 1 In order to talk about trading, exchange, and hybridity, there has to be
some stable notion of the entities that are engaged in that trade. But could such cul-
tures really be pure and completely stable? What differentiates a pure culture from a
hybrid culture; what, in fact, is a scientific subculture?
We will never get anywhere with a too-rigid notion of stability or purity. Your body
is constantly replacing cells, but enough of it remains for it to be possible to identify
and reidentify yourself as the same person. Quasi-stability, not rigidity, is the relevant
criterion: by “quasi-,” I mean that the changes in a given period are small relative to
that which stays roughly the same. We reidentify a university as the “same” even if
it were to pass, as so many have, from being a seminary through periods of being a
sheltered, private teaching institution to being an outwardly looking research and
teaching facility. Scientific practices can and do form subcultures—and the question
is, similarly: Do the commonalities across periods of change hold stable enough to
merit reidentification? That said, all of science remains in flux, and every tired attempt
30 Peter Galison
to grab hold of the necessary and sufficient criteria for scientificity has failed in one
way or another. Quantification? Much of morphological biology isn’t quantitative.
Prediction? Most of evolutionary theory would utterly fail that test. Experimental?
String theory, despite its remarkable contributions to mathematics and to a theoretical
elucidation of black holes and field theory certainly is not that. Explanatory? If you
demanded persuasive explications of the action of many proven life-saving drugs
before you took them, you would die in the waiting room. No, we are getting nowhere
if we start with the idea that there is a pure, stable, transcendental “nature” of physics,
chemistry, biomedicine, or mathematics. What we have are quasi-stable scientific
subcultures (roughly shared ways of handling practices with their attendant values,
symbols, and meanings). Above all, we need to know how these scientific subcultures
connect to each other, to the surrounding world, and to change.
With How Experiments End and Image and Logic done, I had a pretty good idea how,
following experiments and instruments, one could track incipient trading zones
between the laboratory and wider technical cultures. For example, Luis Alvarez knew
how to flip a hydrogen liquefier from producing hydrogen for Atomic Energy Com-
mission H-bombs at the Eniwetok atoll to making hydrogen for AEC bubble chambers
at Berkeley. In the material and work exchanges between the civilian and military
sectors, one can see a great deal: movement of expertise, personnel, materiel, and
funding. We can see, quite dramatically, how the culture of nuclear physics research
took on new forms—a new scale of work more like a factory than a cottage industry,
with semiworks, hierarchical administration, and a new division of labor between
builders and users. Physicists developed new kinds of demonstrations using computer-
aided analyses and simulations alongside a novel scientific-engineering identity for
the practitioners.
But how did exchange work in the dominion of theory? To answer this question,
I started with the very paradigm of theoretical science, Albert Einstein and his work
on relativity theory. It never had made much sense to me that the young physicist
was working in a patent office on new electromagnetic devices fifty or sixty hours a
week and on the foundations of electromagnetism and that the two had nothing
whatsoever to do with one another. One day, I was idly staring at a line of electric
clocks in a European railway station. They seemed to be quite well aligned—but when
I noticed that even their second hands were marching in lockstep, it was clear that
these were not just good clocks; they were electrically synchronized. I wondered: could
Einstein have been thinking of real, not just imaginary, synchronized railway clocks?
The literature on patents at this time bore out my speculation: in 1904–1905, there
was a spike in the already intense interest among Swiss clockmakers in taking out
patents on synchronization mechanisms for railroad clocks. More than that, synchro-
nized clocks had become a hallmark of urban modernity, useful for long-distance stock
exchange trades, but also a shining testament to the pace and vigor of city life. Patents
on electric and electrosynchronized clocks would have landed on Einstein’s and his
colleagues’ desks—and, in mid-May 1905, he used train clock synchronization as the
“metaphor” by which he defined simultaneity in his relativity paper. But to under-
stand the nature of the binding ties between the literal-practical and the metaphorical-
theoretical, I wanted to see this played out elsewhere. I wondered who else would
have been worried about both the technology of time coordination and the physics
of simultaneity?
Henri Poincaré was the obvious candidate, though I started in the wrong place,
looking for ways in which his teaching at the École Professionnelle Supérieure des
Postes et Télégraphes might have bound time signal exchange and his theoretical work
on simultaneity. That was wrong. Instead, a much more fruitful line of inquiry opened
up out of the seemingly “pure” metaphor he used to explain the procedural-material
way in which simultaneity needed to be specified. Einstein had launched his critique
of absolute simultaneity by reasoning about train clocks. Poincaré began his The
Measure of Time (1898) with an allusion to two telegraphers sending signals back and
forth to establish longitude differences. As it turned out, Poincaré was very much
involved with the Paris Bureau of Longitude—corresponding with his British counter-
parts, struggling to sort out technical aspects of the telegraphic exchanges across the
Channel—exchanges designed precisely to sort out simultaneity to a few thousandths
of a second. More: Poincaré had been the spokesman for a dangerous, multiyear lon-
gitude expedition to the Andes; he had even served an important stint as president
of the Paris Bureau of Longitude.
Here, in the procedures of simultaneity, was a trading zone with theoretical physics.
The statement S: “Two clocks A and B are synchronized, and simultaneity defined,
when a back-and-forth signal taking time 2T is exchanged from A to B and back;
assuming the one-way signal takes just time T, when A sends a signal at her noon to
B, B sets his clock to noon plus T when he gets it.” Let’s be specific. Say I send you a
signal at noon, and suppose that it takes two millionths of a second to go back and
forth. Then, when you get my noon signal, you set your clock to noon plus one mil-
lionth of a second. What is statement S? Is S “truly” a physics statement and only
derivatively one from the engineering effort to map the world? (It is not hard to find
versions of S featured prominently in many of Poincaré’s physics publications.) Or is
S really a statement from engineering and only derivatively one from physics? (It was
actually a procedure used every day by the French military geographers.) Or is S in
32 Peter Galison
the first instance philosophical? (It certainly showed up in Poincaré’s epistemological

writings on the nature of time.)
My view is that attempts to make one subculture the basis, the unmoved prime
mover, are doomed to failure. Instead, we would do much better to think about the
arena of simultaneity and synchronization as one area illuminated, as if by intersect-
ing searchlights, by all three sets of practices—physics, engineering, philosophy.
Though deployed differently in each of these discourses, statement S—the procedural
definition of simultaneity—was part of all three. This multiplicity of partially overlap-
ping time talk is why such ideas shone so luminously. Once he had the clock coordi-
nation procedure, Poincaré jumped back and forth, month to month, addressing
longitude finders, physicists, and philosophers—sometimes changing register within
a few weeks.
S should not be thought of as bolted to a granite base above which floats a deriva-
tive and ephemeral superstructure. Instead, S stands in the intersection of three roads
of practice. I choose the idea of an intersection quite deliberately—precisely to avoid
assuming that there is a fixed starting point, a one true source that other domains of
the sociotechnical world uniformly “appropriate,” “reflect,” or “translate.” An inter-
section is awkwardly said to be “in” this or that of its defining paths: Times Square is
not on Broadway any more than it is on Seventh Avenue or on Forty-second Street.
No, the whole point of an intersection is that it lies in all the roads that cross. Trading
zones are such intersections of discursive and material practice, partially—but not
completely—shared.
The key concept here is incomplete coordination. I hand you a salt shaker and in
exchange you pass to me a statuette. We may agree to the trade—we do not in any
sense have to agree to the ultimate use, signification, or even further exchange value
of the objects given. The only thing we have to come to accord about is their exchange-
ability. While for me the statuette may be a religious object, for you it could be a
purely aesthetic or functional one—on this we do not have to agree. We strip away
meaning and memory when we pass the object to a trading zone. As linguists have
long known, this cutting down, this regularization of our symbolic systems is some-
thing at which humans seem to be quite good. And exactly that creation of regularized
interactions and partially interpreted objects marks the trading zones of science.
Back in the 1960s, two outstanding particle theorists, James Bjorken and Sidney
Drell, decided to write a textbook on quantum field theory, and proceeded in two
parts. In the first, they addressed “our experimental colleagues and students interested
in particle physics” (Bjorken and Drell 1964). The goal was to transform “quantitative
calculation, analysis and understanding of Feynman graphs into a bag of tricks” useful
to a larger group of theorists than those fully in command of quantum field theory.
The second volume (Bjorken and Drell 1965) would then fill in the gaps, prove theo-
rems, and explore the regions where ordinary Feynman diagrams could not go. This
move toward regularized, rule-governed procedures and away from the surrounding
or underlying theoretical structure marks a dramatic shift in register. Out-talk is
marked by the connections of rules of calculation to patterns of observation; in-talk
moves among the concepts, exploring relationships and demonstrating systematic
properties within the theory itself.
Does that mean Bjorken and Drell’s second volume for experimenters is simpler,
derivative, less important than the first? Not at all. In fact, some of the detailed cal-
culations of specific scattering processes are significantly more elaborate than the
proofs that follow in the more theoretical volume. But the out-talk volume written
for experimentalists was, without any doubt, more procedural. To find out how likely
X is to happen, draw diagrams Y and assemble a mathematical expression from those
diagrams using rules Z. Integrate and solve.
Linguists are well aware that there seems to be a cross-cultural capacity in language
to be able to switch registers, to shift to more regularized uses of syntax, semantics,
and phonetics. We can and do quite deliberately (in English) switch to subject/verb/
object syntactic constructions, and drop embedded dependent clauses; we can restrict
vocabulary (limit the lexical structure); and (in phonetics) move from complex vowel
strings to a highly regular CVCV construction, in which consonants (C) and vowels
(V) alternate. This flattening of exceptional constructions often characterizes “out-
group” communication, for example to a new language learner. It is precisely this
change in register (regularization) that characterizes the difference between the two
volumes of the Bjorken and Drell textbook on quantum field theory. They wrote
volume one (explicitly) for experimentalists, while they produced volume two (explic-
itly) for theorists. Experimentalists get a form of out-talk, a version filled with very
elaborate uses of quite difficult applications of Feynman diagrams, but stripped of the
talk about Feynman diagrams that explores exceptions, mathematical difficulties,
internal structure, proofs, theoretical analogies. One final, fascinating bit: Bjorken and
Drell suggested that the out-talk (the Feynman rules) might well outlive the in-talk
(field theory)—that the diagrams may become the foundation, with the field theory
nothing but a “superstructure.”
The point of emphasizing the power of what goes on in the trading zone is that
the trading zone is not “mere mortar” between the solidity of bricks. What is exchange
work today may well become the disciplinary pillars of tomorrow: science is forever
in flux, not just in its results but in the contours of its disciplines. Nanoscience began
34 Peter Galison
as an interdisciplinary initiative, but by the early twenty-first century it loomed as a

major continent in the map of the sciences. In the labs of nanotechnology, the atomic
physicist, surface chemist, electrical engineer, and molecular biologist make common
cause. And as they seek to construct objects a billionth of a meter long, their refrain
is: Leave the inessential behind; bring the necessary to work.
We regularize as well when deploying material means in action. If you are teaching
a beginner to fly, you make every landing the same; then, only gradually, you intro-
duce the myriad of particular exceptions to the rules for soft field, short field, and
short/soft field landings, for example. We individuate and modularize concepts, sepa-
rating them from their original multiple and interconnected functions. Like words,
phrases, propositions, and arguments, objects also perform many functions simultane-
ously: a bicycle wheel rolls (like a log), stabilizes (like a gyroscope), and maintains
rigidity (like a doubled arch). We are as capable of stripping down these multiplicities
in material means and actions as we are in “strictly linguistic,” symbolic, or diagram-
matic ones. In trading zone science, the disciplines themselves are relentlessly, rest-
lessly shifting shape.
Objection 2 Neoliberal Reductivism The very idea of trading or exchange presupposes

an underlying notion of money, with all the economic assumptions that implies.
Indeed, by its very nature, any model that includes trade reduces knowledge making
to money making, and so is reductive, transhistorical, and transcultural. Trade imposes
ideas of profit, universal valuation, and divisibility, and, worse, assumes a calculated
rationality of self-interest. Isn’t a trading zone at root a free-trade-zone view of science
that, in the end, amounts to a misfired, neoliberal attempt to be universal—a neoclas-
sical economic theory extended too far?
First, we know from a raft of work in anthropological economics that the Western
mode of handling money within a market economy is by no means universal. So the
idea that any form of exchange presupposes an underlying currency, or that money
presupposes a single form of rationality, simply will not hold water. For example,
Stephen Gudeman (2001) argues that there is always a tension between mutual or
community exchange and market exchange. In his view (reaching back through Marx
and Aristotle), the search for profit is by no means universal in every economic forma-
tion. Profit as a desirable outcome of exchange is, in fact, the result of quite particular
forms of work and life.
More generally, cultures produce many ways to exchange goods. In some contexts,
Gudeman reports, a twentieth-century peasant community in Panama found the idea
of making a profit fully unrecognizable. When they did see profit making—through
their contact with outside traders—they found this form of buying and selling almost
incomprehensible. Even the presence of money itself in the form of coins and bills
does not guarantee a particular stance toward it. Another anthropologist, Michael
Taussig (1980), strikingly showed that there were culturally specific groups for whom
the peso could be blessed—and in such a way that, after a purchase, it would eventu-
ally return. As the blessed peso shows, even if money were present in all forms of
trading (which it is not), money alone would imply neither a universal rationality of
how money is used nor an acceptance of the properties of money that would find a
recognizable description in a standard Western economics textbook.
Second, from a long tradition of work within anthropology, going back at least
as far as Marcel Mauss, we know that exchange relations can be of many types. There
are gifts, as Mauss made clear, that can carry varying degrees of symbolic-personal
baggage, incur obligations on the part of the recipient, and more generally function
outside a simple model of neutral objects that pass from one person’s possession to
another. For example, there are general, or unreciprocated, gifts (e.g., from parents
to children); there are direct forms of bilateral exchange that demand reciprocity
(barter is one type of such an exchange but not the only type); there are circular
exchanges (where X may give to Y, Y to Z, and so on … until somehow, and maybe
much later, someone gives back to X). In her Beamtimes and Lifetimes (1988), Sharon
Traweek used the anthropologist’s contrasting notions of circular and bilateral
exchange to analyze the movement of postdoctoral researchers in particle physics
from lab to lab, and models their circulation on the nonreciprocal exchange of women
among groups.
Does the circulation of postdoctoral researchers among laboratories presuppose a
specifically monetary logic of exchange involving profit? Of course not. Trade and
exchange form a broad genus of which Western, neutral, monetized, storable, divisible
cash is but a single, very particular species. In fact, as Thorstein Veblen pointed out
long ago (Veblen 1915, ch. 3), it is a form of “derangement” to extend our conception
of a single form of monetarized exchange everywhere—to run away with metaphor,
to allow financial capital, for example, to be confused with industrially productive
capital.
If we are going to avoid such derangements of overgeneralization, we must stay
focused on the specific kind of exchanges relevant to exchange languages in the
scientific-technical trading zone. At root, the relevant aspect of exchange is this: what
an object means to me when I give it to you may very well not be what you, as the
recipient, understand that object to connote. What matters is coordination, not a full-
fledged agreement about signification. I hand you a crystal, you hand me a flute. All
36 Peter Galison
we need to know in that moment is that we agree to exchange—not the structure of

the crystal or flute, not their origin, meaning, uses, or provenance. Nothing in this
swap requires a reference explicitly or, for that matter, implicitly to money as a com-
mensurable entity, to a universal instrument of value, or to a universal logic. It is the
possibility of this relative superficiality—the possibility of a thin description that inter-
ests me in the trading zone. It is thin insofar as we do not need to refer to some
universal currency of rationality or value. And thin in a second sense: we can bypass
the presupposition that there is any agreement among the people exchanging things
about the full signification (or thick description) of the objects exchanged.
In the old battle between logical positivism and antipositivism, we have a fatal
recapitulation of the struggles between Enlightenment ideas and romanticism. The
logical positivists, who modeled themselves quite explicitly on the Encyclopedists of
the eighteenth century, desperately wanted a universal common divisor, a language
of science (Frege-Russell logic plus experiential protocol statements) that would cut
across theories, places, peoples, and times. The antipositivists from Kuhn on down
wanted worlds apart, more akin to Boasian cultures, the legacy of nations, each with
its own incommensurable worldview. Frameworks, paradigms, programs—each aimed
to capture a scientific world that stood on its own, that could be judged only accord-
ing to its own terms, that denied absolutely the Enlightenment characteristica
universalis.
The picture of scientific cultures and subcultures tied to each other and to technical
and other cultures of the wider world fits neither the Enlightenment nor the romantic
view. The structure and content of specific trading zones are by no means universal.
The fragmentary, halting attempt to build up structure between biology and chemistry
is not automatically a key to cracking the relation between biologists and physicists.
There is no protocol sentence or one-size-fits-all logic of combination. But at the same
time, while the trading zone picture cracks the perfect autonomy of the romantic
paradigm, its splendid isolation is not so secure. Instead of languages that are purely
“Newtonian” or “Einsteinian,” we have a hybrid mix of local structures bridging the
two: theories with an ether and no way to detect it, experiments that allowed local
comparison of electron flight in electromagnetic fields, theories with ether and no
particles, theories with particles and no ether—and much in heaven and earth besides.
Trade focuses on coordinated, local actions, enabled by the thinness of interpreta-
tion rather than the thickness of consensus. Thin description is precisely what makes
it possible for the experimentalist and the theorist to communicate, albeit in a register
that by no means captures the full world of either, let alone both. Thinness is what
makes it possible for the surface chemist to work with the atomic physicist, the virolo-
gist with the electrical engineer, the computer scientist with the molecular geneticist.
The theorist does not have to probe the myriad laboratory procedures that lie behind
the experimentalist’s confidence that liquid hydrogen has particular thermodynamic
properties; and the experimentalist does not need to know the full mathematical-
physical reasoning backing the theories that lie behind a calculation. What they need
is consensus in a restricted zone, a zone where coordination is good enough.
Objection 3 Power and Diffusion The very nature of trading seems to presuppose a
voluntary agreement between equals, as if power differences did not exist. Doesn’t
this skew the very nature of exchange between different scientific and technological
actors? Doesn’t a power asymmetry make an analysis of contact in a trading zone
impossible when power imbalances are in play? What happens when power is maxi-
mally unbalanced, when there is a simple domination, restructuring the subordinate
in the image of the powerful?
There was a time when imperial history aimed to show that the British or Americans
or French were the affecting, unaffected masters of the world. Their languages and
cultures and economies were supposed to supplant those of the locals with no residue.
Imperialists were viewed as civilizing forces by some and as despoilers by others, but
whether they were seen as sanctifiers or sackers of the castle, historians tended to agree
on this: Delhi was reshaped by London, Dakar by Paris, Samoa by Washington—but
not for a moment were the imperials themselves seen as being reshaped by their
encounter with the conquered. The problem is that none of this story of one-way
cultural imperialism held up much past the mid-1960s. No good history of the last
half-century tells of French impressionism without including the encounter with
Japanese prints and stamps or African masks; no analysis of the origins of American
jazz or rock and roll can be composed without including the musical culture of West
Africa or modern African-American music history.
Returning to science, I’d like to look at the heartland of the purest pure physics:
quantum electrodynamics. The powerful and prestigious theorists—the young
American Julian Schwinger and his opposite number in Japan, Sin-Itiro Tomanaga,
who would each later win a Nobel Prize—surely spread their knowledge, imposed their
views, “disseminated,” “radiated,” “multiplied,” “diffused” their knowledge down
one-way channels issuing from the center. It seems to be true, in this case, that the
center transmits, while the periphery (more or less properly) receives.
But dig a bit. It turns out that Schwinger and Tomanaga reformulated the founda-
tion of physics after World War II. During the war, they had fought on opposite sides,
each working with their respective radio engineers, each side at first having cobbled
38 Peter Galison
together radar sets, later groping their way in the new domain of microwave engineer-
ing and industrial production. Before the war, radio engineers, who were low in
prestige compared with theoretical physicists, wouldn’t have shown up, at least in the
United States, at the highest levels of the most prestigious universities. But the war
shifted the relation of physicists and engineers during the years spent reformulating
electrodynamics so that the radio engineers could use it to fight the war. The high-
flying physicists began to see their own endeavor in the image of radio engineers:
from 1946 to 1948, black-box input-output analysis and relations, “effective” circuit
elements, and modular calculation strategies began to show up inside the heartland
of the high-born theory. Both Schwinger and Tomanaga testified that their work on
wartime radar had been important for quantum electrodynamics.
Maybe the purest of the pure is thus not quite so far as we imagine from the black-
speckled microwave transmitter mounted in a B-29. For Schwinger and the radio
engineers, and indeed for the vast majority of the American and British panoply of
physical sciences, the propellant was war. With the Blitz pounding London, there was
nothing abstract about the Nazi threat—there would be an effective radar system, or
the war would be lost. External forces—war, economics, natural disaster—can drive
participants into exchange. The radar engineers had no idea how to produce effective
circuits without the help of the physicists. And the physicists were in no position to
design the apparatus without the experience of the radio engineers.
Intriguingly, here we have an example where the trade—the coordinated exchange
between electrodynamic theory (by theoretical physicists) and very pragmatic micro-
wave circuit design (by radio engineers)—reshaped high theory. It did so not by
importing microwave resonators or antennae directly into quantum electrodynamics.
Instead, the syntax, so to speak, of a laboratory science—the characteristic rules of
manipulation—got taken up by the physicists. As an example: radio engineers had a
way of analyzing problems that required the reduction of each component to an
equivalent circuit (a circuit that had the same input-output relations but was physi-
cally much simpler; the engineer ignored physical details that had no importance to
the output). Immersed in such design problems when he was assigned the wartime
task of producing equivalent circuits for microwave devices, Schwinger learned to
calculate things by ignoring everything that was not essential to the task of relating
input to output.
Schwinger learned, for example, to ignore those aspects of particle collisions that
were not important to the final state of the system. This systematic stance of the radio
engineer—to focus on the input-output relations and ignore physical complications
that do not affect the final state—became the basis for Schwinger’s take on the devel-
opment of the physicist’s most abstract achievement: renormalization theory, which

showed how to get finite, precise predictions from the theory of quantum electrody-
namics. Here we have a striking example of a trade conducted between communities
of very different prestige and authority: physicists from the most powerful universities,
working in a radar program that they controlled, trading with radio engineers.
From examples like the joint work of radio engineers and theoretical physicists, a
new question arises: What characterizes the forms of technical exchange that take
place under such conditions of inequality? This, as it turns out, is a question that
arises in nonscientific interlanguages. Some anthropological linguists have argued that
the subordinate group often donates syntax, while the superordinate group provides
lexical or referential structure. I suspect that something similar went on between the
physicists and the engineers: calculational strategies were from the engineers, terms
from the physicists. While this pattern may not be universal, it is suggestive. At the
very least such examples prompt a set of questions: In instances of unequal exchanges
between scientific-technical subcultures, what precisely does make it to the interlan-
guage from each side? It is a question that cannot even arise if we stop our analysis
with proclamations about “interdisciplinarity,” “collaboration,” or “symbiosis.” Those
terms point at the problem; all the interest, in my view, lies in unpacking what the
nature of this coordination is and how it evolves over time.
We can then ask the reciprocal question: What happens at the other extreme—
when the groups involved in trade are more or less equal, rather than utterly disparate
in their prestige and authority? What drives exchange in the absence of command
structure (of the kind a government exercises in total war)? Though I can give only
the most schematic of sketches here, the following is an example from the early 1990s,
when mathematicians (more specifically, algebraic geometers) confronted physicists
(string theorists). Both groups were quite prestigious within their professions, both
thoroughly abstract; neither could be said to have had the upper hand.
Since the late nineteenth century, mathematicians had struggled to count the
number of independent curves of a given type that could be drawn on a surface. Over
decades, with Herculean efforts, the “enumerative geometers” worked out the first of
several such problems, coming up with 2,875 curves of type one (n1). Then, in 1986,
came a triumph: Harvard’s Joe Harris managed to find the second-degree curves
(conics), and reckoned their number to be n2 = 609,250. Moving up to curves of the
third degree promised to be painful beyond measure, but two Norwegian mathemati-
cians set themselves the task with the help of a cheap computer. By 1990, they had
a result. At the same time, using utterly different methods having to do with the scat-
tering of one string by another, a group of physicists claimed to be able to resolve
40 Peter Galison
such problems as the calculation of n2. When the mathematicians tried to flick them
away, the physicists offered to provide n3 … and n4, n5, and even n10 … in a matter
of hours.
Each group found almost everything about the other’s approach incomprehensible.
They used different methods, they had different objects and a different vocabulary.
But they knew one thing: they clashed—over a number. For now matter how you
sliced it, the two results were incompatible:
Physicists: n3 = 317,206,375
Mathematicians: n3 = 2,682,549,425
Not much later, the Norwegian mathematicians discovered an error in their computer
program, and they too got n3 = 317,206,375. From that moment in 1990, the two
sides knew they had to sort out what the other side was doing. The mathematicians
saw that the physicists had, by methods utterly unknown and indeed incomprehen-
sible to them, found an easy road to results that the mathematicians themselves had
struggled to achieve for decades. The physicists now understood that their hunt for
an account of string theory collisions had taken them into mathematics—and they
wanted to understand the strange spaces in which the strings would live. For a brief
moment, the two groups shared what amounted to the world’s smallest imaginable
trading zone: a single number, n3.
Bit by bit, what began as a punctiform trading zone became much more. Jointly
authored papers, conferences, and graduate programs began to emerge; and in concert,
but not without tension, mathematicians and physicists composed a growing but still
restricted vocabulary and set of procedures. Within a few years, they were debating
the virtues of training a new generation of scientists who could move back and forth
between mathematics and physics, exploiting not only the concepts and methods but
also the intuitions of both. This was a trading zone propelled not by external demands
of the state, but instead by the separate—and quite different—ambitions of the two
sides.
If one is content to label work between scientific subcultures as “interdisciplinary,”
questions remain that are utterly obscured. Of course we know there is collaboration—
that is what we want to understand. To tackle the joint workings of different groups
by referring to a label is not much help. It reminds me of Molière’s quack who explains
the sleep-inducing power of opium as being its virtus dormitiva. What we need is a
much more interesting and effective active ingredient than “virtus dormitiva”—
instead a way of approaching joint work that parses what comes with what, and how
ways of speaking, calculating, and building are coordinated.
Objection 4 Language and Materiality Within the trading zone, exchange languages—
scientific jargons, pidgins, and creoles—are supposed to structure the nature of what
is handed back and forth. Isn’t this use of language just a metaphor from linguistics,
an unrelated field? Worse, if science really is nothing but linguistic, where does that
leave us with material objects? My point in writing How Experiments End and Image
and Logic was to reintroduce the materiality of argumentation; yet doesn’t exchange-
language talk eviscerate the materiality we have worked so hard to reinstall in the
study of science?
We have been trying to understand the linguistic face of science for a long time—
certainly since the beginning of modern philosophy of science between the two world
wars. Rudolf Carnap’s (1937) “logical syntax” was directly and explicitly an attempt
to get at the structure of argumentation without buying into what he considered the
inevitable subjective metaphysics of trying to ground the language of science through
the direct and subjective referentiality of statements like “I see red.” Could I ever
really know how someone else experienced blue? Of course not. Instead, relational
structures—locating blue on the spectrum—could be shared (we agree blue is between
violet and red), and this syntactic rather than semantic structure would undergird
objective knowledge. In addition to Carnap’s approach, Otto Neurath’s “physical
thing language” also made language essential, as did their joint insistence on protocol
language as the sine qua non of meaningful talk in science. Could you or could you
not take your scientific claims and express them through such utterances as “smell
ozone 12 noon here”? If so, proceed; if not (as the Vienna Circle claimed), you may
be making noise, but you are not speaking meaningfully. (For more on Carnap,
Neurath, the Vienna Circle, and Quine, see Galison 1997, chs. 1 and 9; and more
generally Giere and Richardson 1996.)
Quine too spoke of theories as languages—and he pointed out that there would
always be more than one way to translate from one theory language to the other.
Though he famously split from Carnap on some issues, Carnap also constantly empha-
sized the multiplicity of languages that could be invoked to express certain structures.
In his most famous work, Der logische Aufbau der Welt (properly translated as “The
Logical Construction of the World”), Carnap underscored the possibility of reexpress-
ing the same structure in different ways: one could start with “my” experience, or one
could start with a more social-collective base. Either way, the relations among proposi-
tions would remain the same.
During the 1960s’ backlash against logical positivism, philosophers of science
changed their account of science dramatically. No longer would protocol utterances
remain the universal language base of science. But the idea that scientific accounts
42 Peter Galison
were, in the end, a form of language, did not perish. Instead, Thomas Kuhn took the
big paradigmatic theories—of heliocentrism, phlogiston chemistry, oxygen chemistry,
classical physics, and relativity (to name but a few)—and considered them as full-on
languages, analogous in their own right to English, French, or German. Kuhn’s thesis
then put a twist on Carnap’s principle of tolerance (or for that matter Quine’s con-
ventionalism). Where Carnap and Quine argued that there would always be more than
one translation, Kuhn shocked readers by claiming that there wouldn’t even be one.
Speakers of “Newtonian,” as it were, could never, without gaps and awkwardness, fully
translate what they had to say about the world into “Einsteinian.” Any such attempt
would fail for the same reason that there are inevitable misfirings between texts in
German and French: there was no adequate translation.
That Kuhnian picture of full-blown but incommensurable languages—languages
intact in themselves but without a common divisor like a protocol language—smoothly
carried over into much of classical 1980s-style science studies. (Indeed, incommensu-
rable languages grounded the methodological and philosophical commitment to rela-
tivism, and 1980s science-and-technology-studies relativism cast itself as the polar and
only alternative to the putatively referential claims of a naïve realism.) Even in the
rather distant frame of philosophy, Michel Foucault took science and divided it into
epistemes, each of which was marked by a particular logic by which statements
(énoncés), rather than true/false propositions, could be ordered.
Throughout the decades from the 1920s through the 1970s, science was thus for-
mulated and reformulated as a kind of language. Debates raged about the kind of
language it might be, of course, but that it was a language, a form of structured com-
munication, was not really debated. Crucially, the language of science was always a
global one, invariably analogized to a snapshot, unchanging and global, of a contem-
porary French, English, or German.
My original problem—the problem that drove me to the idea of trading zones and
scientific exchange languages in the first place—was my frustration in trying to join
a local picture of practices with this fixed, global idea of language. The two clashed.
By contrast, interlanguages are exactly characterized by their change over time and
by their locality—exactly what one needs in order to talk about scientific language in
the context of a shifting set of laboratory or blackboard practices.
Trading zone languages can be quite heterogeneous: they are sometimes nothing
but a few terms held in common, a bare scientific jargon. As we saw earlier, in the
clash between the string physicists and the mathematicians, they were, at the outset,
at the very minimal limit—all they had in common was the (disputed) number of
curves on a certain surface. Here was a trading zone with nothing in common but the
number of curves of degree three. That particular point zone eventually did expand,
hugely—into a much wider world of mathematical-physical discourse that trans-
formed both “parent” fields (string theory and algebraic geometry). Other trading
zones carried over from a scientific jargon to a scientific pidgin to a scientific creole—
think of biochemistry, which is now quite capacious enough to “grow up in” scientifi-
cally. But there is absolutely no teleological guarantee. Not every jargon gets developed
into a pidgin; not every scientific pidgin molts into a creole in full bloom. In the
nonscientific world, examples abound of pidgins that froze and died: some of the
Korean-English pidgins forged during the Korean War of the early 1950s simply
vanished when the firing stopped. As another example, iatrogenics was a science its
creators hoped would join Newtonian mechanics and physiology. It perished without
a trace after the eighteenth century. So, more or less, did “neutronics,” an interdisci-
plinary field that nuclear scientists, engineers, and health physicists hoped would
flourish after 1945.
Sometimes the language of science does read, quite literally, as language: proposi-
tions, statements, observations, hypotheses, and conditionals are all recognizably
linguistic even if technical in scope. But at other times practices do not necessarily
form linguistic objects, in a strict sense. Diagrams and symbols, for example, have
their own combinatorial logic. We are used to talking about the “language of math-
ematics,” and for good reason. I am interested in language in an expanded sense that
would embrace such symbol languages—whether computer codes, abstract algebra,
formal logic, or the calculations of quantum physics. Each carries with it its own form
of syntax, its own rules of simplification, generalization, and composition. Similar,
though perhaps less familiar, are languages formulated in ways that make use directly
of spatial or topological relationships—electronic schematics, group-theoretical Dynkin
diagrams, Minkowski space-time diagrams, Feynman diagrams (on the latter, David
Kaiser’s work on the piecewise transport of Feynman diagrams is central).
Diagrams too have their rules of manipulation. Reasoning with them does not
necessarily require constantly returning to words or even algebra. Indeed, that’s why
they are so useful: spatial arrangements suggest variations and allow manipulations
without translating into another idiom such as words. (Write down a Feynman diagram
for a particular scattering process and a physicist might say, “What about this
diagram?,” and modify the first by flipping a dotted line but preserving the same
number of vertices, for example.) Moving from the manipulation of electrical diagrams
to the manipulation of circuit elements themselves is not such a big jump—in fact,
there are machines that take diagrammatic representations and produce the circuits.
My view is that the regularized, rule-governed procedures that manipulate material or
44 Peter Galison
symbolic objects are also a form of language, and it becomes entirely appropriate to
speak of wordless jargons, wordless pidgins, and wordless creoles.
In fact, we know perfectly well that such objects move back and forth—every day
we pass musical scores, mathematical symbols, and electrical circuits back and forth
between people who speak different natural languages. So it is within science—
physicists of different theoretical persuasion can view a bubble chamber image and
still find a thin description upon which both can agree: “that’s an electron, hit by a
neutrino, scattering and emitting a photon that becomes an electron-positron pair.”
Or think of the monumental efforts that have been made to produce anatomical,
astronomical, or neurohistological atlases. Throughout the nineteenth century, these
were produced in vast numbers precisely to work in a visual register free of detailed
sectarian interpretation. Such atlases did not need to be simple or peripheral—but
they did need to address an audience outside this or that tendency within the field.
Like out-talk by speakers of a language, doctors or astronomers could produce images
and objects open beyond the originating culture.
This set of thoughts returns us to the root idea of the whole scheme of trading
zones: it is possible to share a local understanding of an entity without sharing the full
apparatus of meanings, symbols, and values in which each of us might embed it.
Images, symbol systems, calculational and diagrammatic schemes—even complex
objects—could be part of a generalized notion of language that is far from “just words.”
Indeed, language, as I want to use it, is a regular yet flexible apparatus that may take
many forms, from the recognized, everyday “natural world languages” to the myriad,
systematic registers in which we communicate.
Objection 5 Applicability For a concept to be useful, it must have limits. But if every-
thing is always hybrid, if every situation admits of a trading zone, then isn’t this
concept just a restatement of what we already know? What isn’t hybrid; where does
exchange not take place? In other words: What are the useful limits of the concept of
the trading zone?
Exchange involves coordination between scientific-technical cultures. These cul-

tures are specified by practices that pick out a certain quasi-stable configuration of
practices—and the meanings, values, and symbols linked to the practices. But the
necessary condition for a trading zone is that practices (and their interpretations) tend
to travel in packs rather than along arbitrarily combined trajectories. These “packs”
might be a set of affiliated experimental procedures in organic chemistry in the early
twentieth century; or they might be the mathematical toolkit of the quantum physi-
cist in the 1930s. Here they are tactics, there they are strategies—but also regulative
values (what counts for mathematicians as well defined; what counts as a proof).
Together, the assembly of practices, values, and meanings do more than simply pick
out a problem-solving mechanism; they also set out the contours of scientific identity,
defining what it means to say, “I am an experimental high-energy physicist,” or “I am
a theoretical biologist.”
In the 1980s, many particle theorists viewed high-end mathematics (of the type
string theorists practiced) with reserve, if not outright moral suspicion. They judged
the ins and outs of Calabi-Yau spaces to be too fancy to have purchase on reality, not
close enough to lab results. By contrast, nuclear physics and much of atomic physics
seemed to those same 1980s theorists to be insufficiently theorized—too close to the
measurable, too liberal in their acceptance of heuristic, phenomenological, and partial
models, too weakly mathematized. Together, skills and stances offered the late-
twentieth-century particle theory practitioners a way of looking at their corner of
science and at what they stood for as scientists. Around certain practices came not
only the meaning and symbols, but also the virtues and sins that gave a moral struc-
ture to this cut through scientific life. More generally, practices—along with the values
associated with meanings and symbols—offer the defining attributes of scientific or
technical subcultures.
Should we characterize any set of embodied practices as a subculture? The question
is an empirical one. Is there enough regularity, enough covariance within a given set
of practices, to merit our picking out that regularity for attention? We have to be
prepared sometimes for the answer to be “no.” If there is enough regularity to justify
speaking of quasi-stable subcultures in contact with one another, then, and only then,
is the trading zone idea useful, because it is then that the thinness of the exchange
proves valuable—in contrast to the thickness of the established cultures.
For emphasis: the trading zone concept is not always applicable. Indeed, we know
that many sciences—physics included—at some historical moments do not have this
particular partition: experimenting, theorizing, and instrument making. Certainly
such a division was not a commonplace in the time of Galileo. It would be equally
distorting to split Gregor Mendel’s work on segregation and assortment into a “theo-
retical” and “experimental” biology. In physics, William Thomson (later Lord Kelvin)
cannot be thought of as either an experimenter or a theorist. But we can go farther.
The right question to ask (about Thomson’s work, for example) is this: Does Thomson,
when theorizing, participate in a discourse (an ordered set of practices, whether
linguistic, symbolic, or physical) that forms a roughly covarying set? Thomson strug-
gled to relate the structure of atoms to the nature of ethereal vortices, the generation
of smoke rings, and the theories of knots. In that case, he clearly was working in
46 Peter Galison
different practice sets—knot theory, for example, linked him to a group of mathema-
ticians, whereas his broader theory of atoms connected him with other physicists.
The relevant trading zone question is then empirical-historical: What was and wasn’t
shared between the broader atomic theory to which he was committed and his
specific work on knots—what pieces of each were linked, and in what ways?
There is one final class of limits to address—the limits in which the trading zone
concept comes into contact with other work in the understanding of science. (I have
in mind a kind of mathematical limit rather than a limit of validity: 1/√(1 − v2/c2) goes
to 1 when v/c is small but to infinity when v/c approaches unity.) First, consider cases
where the exchanged object is not, in fact, part of an ordered trading language—where
the object stands alone, so to speak, not subject, or at least not importantly subject,
to rules of combination and association in the trading zone. In this instance, one has
objects that sit on the boundary that can be compiled, collected, and used by different
groups. This corresponds to Susan Leigh Star and James Griesemer’s (1989) very useful
notion of “boundary objects,” which they developed to discuss the flora and fauna
collected by California amateurs to document the forms of life in the Golden State.
As they showed, professionals in different fields used these samples in different ways.
Unlike the case, say, of algebra and geometry, out of which algebraic geometry was
formed, or biology and chemistry combining to constitute biochemistry, the archival
flora collections didn’t, in the long run, become a field in their own right, for a col-
lection is not a discipline. Boundary objects might be thought of as a kind of time
slice of a trading language where the lexical lists exist, but our attention is not focused
on the syntax.
A second limit of the trading zone occurs in the limit of an asymptotically large
power difference between the groups trading. We have seen cases where the power
differential was small (as it was with the mathematicians and string theorists); we’ve
seen cases where the domination of one group by the other was fairly significant
(physicists working on radar with radio engineers). But one can imagine instances
where the discrepancy is so enormously huge that essentially no input comes at all
from anywhere but the superordinate group. Given that even slaveholder cultures were
reshaped dramatically by slave culture, it is not clear that such an absolute gap could
exist, but in that limit, one could imagine a scientific laboratory that imposed itself
like a kind of implantation, a colonial outpost that repelled all forms of locally pro-
duced materials, machinery, products, or personnel. In that forceful extension of the
center, one indeed would find the kind of situation that Simon Schaffer (1991) cap-
tures so well with his notion of a “multiplication of contexts,” by which a laboratory
in London or Paris could move its experimental apparatus. Or think of Bruno Latour’s
(1985) important work on the ways the world must be configured to allow the scientist
to “travel very far without ever leaving home,” the material analog of his “immutable
mobiles” (movement without going anywhere). These are indeed the extremes of
power imbalance, the annihilation of locality, the far-limit horizon of cultural contact
in which the superordinate group hugely outweighs the subordinate one.
Finally, there is a third limit point of trading zones. Consider what happens when
a group presents out-talk in the absence of an engaged interlocutor from another,
distinct group. Here, I do not mean the atomic physicist presenting her work within
the nanolaboratory to the surface chemist and virologist. Instead, think of the presen-
tational, reflective, informal work among such practitioners as they walk away from
the laboratory, talk over lunch, or begin to sketch out a paper or presentation. This,
I would argue, is the limit situation that gives rise to what Harry Collins has dubbed
interactional expertise (see, e.g., Collins and Evans 2002). Collins considers an inter-
esting case—the verbal, unmathematical discourse of the gravitational wave physicists.
Interactional expertise is a very interesting notion, one the too-rigid canonical social
studies of knowledge has (in my view) dismissed too quickly. Collins’s neo-Turing
argument is strong and persuasive (the test is this: Can an outsider learn enough
vocabulary and characteristic ways of speaking to simulate a “real” gravitational physi-
cist as long as no mathematics, hardware manipulation, or calculations are involved?).
It seems to me absolutely right that it is possible to learn to interact in such conversa-
tion in a way that is familiar and recognizable to all practitioners—without, in the
process, learning the others’ mathematical or detailed experimental or craft competen-
cies. The capacity to carry on a professional-level, informal conversation about gravity
waves is precisely analogous to particle theorists’ ability to speak in the more regular-
ized, stripped-down manner of out-talk scientific pidgin.
Early-twenty-first-century experimentalists working on the billion-dollar detector
at the Large Hadron Collider at CERN had to converse with the electrical, cryogenic,
and structural engineers. To do so, the experimentalists needed to know how to move
their vocabulary, parameters, and calculation devices into a form that a sophisticated
technical person (who was not a physicist) could grasp. Gone from the experimental-
ists’ local concern were the details of the supersymmetric partner to the photon, the
hypothetical “photino”; gone too were the detailed physics of the supposed Higgs
particle. These physicist-with-engineer discussions were much more about gas
characteristics in the detectors, the failure rate of circuits, and the degree of radiation
hardening against the appalling environment in which the detectors had to live. In
all trading zones, there is always such a shift of register as each of the participating
groups creates an out-talk suitable for communicating with the others.
48 Peter Galison
I take interactional expertise to be the capacity to speak in a specific register, an

acquisition (by an outsider) of a form of pidginized out-talk used by physicists, for
example, to speak with one another with a minimum of ancillary knowledge. Stripped
from these conversations are a wide variety of other kinds of talk: detailed calculations
or proofs from differential geometry on one side, particular issues of materials, instru-
ments construction, experimentation, or engineering on the other. Though it repeats
a theme I keep hitting, it is important to emphasize again that regularized and
stripped-down out-talk is not a lesser version of something else; rather, it is a register
of scientific interaction that is supple and effective in its domain. A creole is not a
poor version of a “parent” language. In precisely the same way, scientific out-talk is
neither identical to the technical language from which it originates, nor a diminished
version of it. The skills of someone versed in interactional expertise represent one
specific register of scientific language. Regularized, demathematized out-talk is then a
third limit of the interlanguage performance within a trading zone.
Trading Zones: Why Now, Where To?
Historians, political scientists, and sociologists regularly think of the Cold War in
terms of international confrontation, domestic political repression, and the arms race.
But we have only touched the surface of how the long war from 1939 to 1989 shaped—
and effectively froze—aspects of the academy. I’d like to suggest, all too briefly, that
the post-Cold War disciplinary map is in a state of intense rearrangement, one unpar-
alleled by any developments since the immediate postwar years of the late 1940s. This
set of shifts has made modes of coordination between and among long-established
fields immediately pressing—contexts in which trading zones are conspicuously
present.
It is easy to think of our universities as highly stable, unchanging fixtures of the
world, so old as to be part of the distant, unremembered past. But the academic forms
we know from the present are much more recent than the antique founding statues
and plaques that adorn the university gates. The world of internationally connected
science, liberally funded by national agencies and open to an increasingly diverse
population of students and faculty, is a creation of the years just after World War II.
So too are the system of national laboratories, competition for contract funding, and
the construction of government-owned and corporate-operated laboratory facilities.
In the United States, the Atomic Energy Commission, the National Science Founda-
tion, the bulk of the National Institutes of Health—just to name a few—also rose in
the shadow of the Cold War. Besides these large institutions, the departments of
universities achieved a new kind of fixity. True, atomic physics gave way to nuclear
physics, nuclear physics yielded to nucleonic structure studies, and the interior reso-
nances of protons and neutrons eventually yielded to particle physics. But physics
from the 1920s through the 1970s remained recognizably physics—the basic courses,
the ethos of training, the divisions between theory and experiment, the prestige
hierarchy of pure over applied work.
Over the course of the Cold War, the essential integrity of the basic departmental
division of knowledge stayed in place, even if new departments would sometimes
appear, such as computer science and biochemistry. Since the fall of the Soviet Union,
some of this fixity has been eroded. New flows of funding bolster different kinds of
research—startups, intellectual property, venture capital—and all have blurred the
lines between the pure and applied sciences. Cryptology went from being a concern
of the national security state to a Web-based industry. Nanoscientific groups often
maintain two Web sites, one for their academic work, another for the corresponding
startup. Universities encourage and even participate as stakeholders in the acquisition
of patents and their deployment in new ventures. All of this means that an increasing
number of students emerge from their doctoral studies with a very different experience
of disciplinary formation. In many arenas of nanoscience, wherever one began—
atomic physics, surface chemistry, electrical engineering—collaborators come from all
three domains. Clean rooms, visualization facilities, and fabrication devices are all
shared. Joint appointments have become increasingly common between physics and
biology, physics and mathematics, physics and chemistry, and so on down the line.
This world of shared space, technique, training, and authorship has increasingly
obviated the spell of the “pure.” The terms “pure physics” or “pure biology” ring false.
More than that, the very idea of pure science—as more prestigious, important,
or consequential than applied science—has lost traction. In the midst of string-
theoretical work, the challenge that an investigation is “not physics, just mathemat-
ics,” doesn’t much move the postdoc—nor does she find herself distressed by the idea
that nanoscale devices are “just engineering,” and therefore not truly physics.
Thus, everyday scientific work already militates for attention to crossover domains,
and away from “pure” languages, theories, and disciplines. But if trading zones have
helped us grapple with pressing boundary work after the Cold War, it can also suggest
ways to look freshly at the other periods in which disciplines were in flux: in that
postwar moment, for example, or in the late nineteenth century.
If the Cold War crystallized disciplinary divisions, its historiography also froze
certain debates so deeply that they seemed to be inevitable features of any historical
inquiry. Historians of art split (in what long seemed an irresolvable struggle) between
50 Peter Galison
the social theory of art, which aimed to tie the making of artworks to social condi-
tions, and formalism, which wanted exactly the opposite—to make the painting
speak on its own by the analysis of color, brushstroke, and other formal features.
Literary historians similarly divided between close readings of texts and situating
them contextually or theoretically. And historians in history departments fractured
into those defending military, diplomatic-political, or high-cultural or intellectual
history (from above) on the one side, and those after social or cultural history (from
below) on the other. What powered these arguments and made them more than
intraprofessional disputes was the long, ferocious battle over Marx. Behind histori-
ography was politics—the hard-fought left/right politics that intensified during the
Cold War.
In the 1950s, 1960s, and 1970s, the history of science spun its own version of
Marx-anti-Marx: internalism versus externalism. Internalists prized the autonomy of
science, its freedom from outside circumstance. Internalism was meant to be a giver
of law to itself, dependent (when done right) on nothing and no one. Externalism,
by contrast, took scientific content to be nothing more than the surface waves caused
by the deeper currents of class, psychology, institutions, or technology. Yet tracking
the flow of space, techniques, language, and standards offers a way to eschew this false
choice between total autonomy and total dependence. Pushing locality all the way
down to scientific techniques, languages, and values offers a way to address practices
as they form among the sciences and between the sciences and the worlds of work
that abut them.
There are many ways to carry on with research on trading zones and interlanguages,
and it would be the worst kind of self-refutation, were I to try to set out what should
or shouldn’t be the “right” way to deploy such concepts. The trading zone offers a set
of tools, not a doctrine. In that spirit, I want to gesture just a bit at the ways that the
idea might have some rather practical consequences. Some, happily, are explored in
this volume. But let me draw an example from elsewhere—from the work of Boyd
Fuller (2006 and forthcoming), whose study of water use battles in California and
Florida exemplifies some of what I have in mind.
Fuller began with conflict. The stakeholders in recent debates over the Everglades
were more than diverse—federal and state regulators, tribal groups, environmentalists,
and agricultural interests “exploded” in some of their early attempts to interact. Their
values were irreconcilable, their desires askew. Fuller showed that these actors neither
subscribed to a common worldview about the meaning and significance of wetland
water supplies nor threw up their hands in despair at the clash of values. Instead,
without abandoning their own deeply held values, the groups were able to establish
terms of negotiation around a delimited set of water management recommendations.

(For additional examples of trading zones involving scientific and technical policy,
see the chapters in this book by Jenkins, by Gorman and Werhane, and by Collins,
Evans, and Gorman.) Over a very broad range of battles from power-generating sta-
tions to fisheries, we have scientists and practitioners struggling to find common—but
restricted—interlanguages. It would be powerful if we could understand more system-
atically why some disputes can be productively advanced through the formation of
delimited trading zones, while other such attempts fail. If we could figure that out,
our understanding might lead us to strategies to encourage positive outcomes. Here,
it seems to me, is a theoretical problem that bears on the most practical side of trading
zone work today.
References
Baran, Paul A., and Paul Sweezy. 1966. Monopoly Capital: An Essay on the American Economic and
Social Order. New York: Monthly Review Press.
Bjorken, James, and Sidney Drell. 1964. Relativistic Quantum Mechanics. New York: McGraw Hill.
Bjorken, James, and Sidney Drell. 1965. Relativistic Quantum Fields. New York: McGraw Hill.
Braverman, Harry. 1974. Labor and Monopoly Capital: The Degradation of Work in the Twentieth
Century. New York: Monthly Review Press.
Carnap, Rudolf. 1937. The Logical Syntax of Language. Trans. Amethe Smeaton. London: Kegan
Paul. Originally published as Logische Syntax der Sprache. Vienna: J. Springer, 1934.
Carnap, Rudolf. 1967. The Logical Structure of the World and Pseudoproblems in Philosophy. Trans.
Rolf A. George. Berkeley: University of California Press. Originally published as Der logische Aufbau
der Welt. Leipzig: Felix Meiner, 1928.
Collins, H. M., and R. J. Evans. 2002. The Third Wave of Science Studies: Studies of Expertise
Dutton, Tim. 1985. Police Motu: Iena Sivarai. Waigani: University of Papua New Guinea Press.
Foucault, Michel. 1972. The Archaeology of Knowledge. Trans. A. M. Sheridan Smith. New York:
Pantheon.
Fuller, Boyd. 2006. Trading Zones: Cooperating for Water Resource and Ecosystem Management
when Stakeholders Have Apparently Irreconcilable Differences. Ph.D. dissertation, MIT.
Fuller, Boyd. Forthcoming. Trading Zones: Cooperating and Still Disagreeing on What Really
Matters. Journal of Planning Education and Research.
Galison, Peter. 1987. How Experiments End. Chicago: University of Chicago Press.
52 Peter Galison
Galison, Peter. 1988. History, Philosophy, and the Central Metaphor. Science in Context 2
(1):197–212.
Chicago Press.
Galison, Peter. 1998. Feynman’s War: Modeling Weapons, Modeling Nature. Studies in History
and Philosophy of Modern Physics 29(B): 391–434.
Galison, Peter. 2003. Einstein’s Clocks, Poincaré’s Maps: Empires of Time. New York: W. W. Norton.
Galison, Peter. 2004. Mirror Symmetry: Values, Persons. In Growing Explanations: Historical
Perspectives on Recent Science, ed. M. N. Wise, 23–66. Durham: Duke University Press.
Galison, Peter, and Lorraine Daston. 2007. Objectivity. Cambridge, MA: Zone Books.
Giere, Ronald N., and Alan W. Richardson. 1996. Origins of Logical Empiricism. Minneapolis:
University of Minnesota Press.
Gudeman, Stephen. 2001. The Anthropology of Economy. Malden, MA: Blackwell Publishing.
Kaiser, David. 2005. Drawing Theories Apart: The Dispersion of Feynman Diagrams in Postwar Physics.
Chicago: University of Chicago Press.
Latour, Bruno. 1985. Science in Action. Cambridge, MA: Harvard University Press.
Mauss, Marcel. 1923–1924. Essai sur le don. Forme et raison de l’échange dans les sociétés
archaïques. L’Année Sociologique.
Sahlins, Marshall. 1972. Stone-Age Economics. New York: Aldine de Gruyter.
Schaffer, Simon. 1991. A Manufactory of Ohms. In Invisible Connections, Instruments, Institutions,

and Science, ed. Robert Bud and Susan Cozzens, 21. Bellingham, WA: SPIE Optical Engineering
Press.
Taussig, Michael. 1980. The Devil and Commodity Fetishism in South America. Chapel Hill: Univer-
sity of North Carolina Press.
Thompson, E. P. 1966. The Making of the English Working Class. New York: Vintage.
Traweek, Sharon. 1988. Beamtimes and Lifetimes. Cambridge, MA: Harvard University Press.
Veblen, Thorstein. 1915. The Theory of Business Enterprise. New York: Scribner.
Wittgenstein, Ludwig. 1953. Philosophical Investigations. Oxford: Blackwell.

4 Interactional Expertise and the Imitation Game
Robert Evans and Harry Collins
Introduction
Interactional expertise provides one solution to the problem of coordination created

by the existence of different cultures. Though it is not the only resolution, it has
particular relevance for social scientists as it justifies their own status as experts. Put
another way, if there was no such thing as interactional expertise, interpretive sociol-
ogy would be impossible unless social scientists more or less completely shared the
physical experiences of those they research. But, to give one counterexample, a crimi-
nologist can succeed without first committing crimes. Indeed, if shared practice were
always a prerequisite for understanding, then not only would social science in general
become impossible, but each of us would live in small private boxes of practice con-
ceptually opaque to our collaborators, family members, and lovers. We call the exper-
tise that bridges distinct practice through a deep sharing of discourse “interactional
expertise.”
Interactional expertise can be acquired by the usual techniques of social science
fieldwork—participant observation and immersion in the discourse of a community.
The strong claim is that immersion in the discourse is just as good as immersion in
the practices, so long as the aim is competence in tasks in which practice is not
required (Collins 2004, 2007; Giles 2006). In a trading zone, a person with interac-
tional expertise could move smoothly between different social groups, “translating”
the concerns of one into the language of the other and vice versa (Ribeiro 2007a,
2007b; Shrager 2007; Collins, Evans, and Gorman, this volume). Such abilities are
what make the division of labor possible.
The idea of interactional expertise is, therefore, important even though, in practice,
separating the linguistic and practical elements of expertise is not easy. In what
follows, we set out the idea in more detail before describing an experimental protocol
for investigating it empirically.
54 Robert Evans and Harry Collins
Periodic Table of Expertises
The idea of interactional expertise is part of a more broad-ranging approach to the

study of expertise known as studies of expertise and experience (SEE). First proposed
in the paper “Third Wave of Science Studies” (Collins and Evans 2002) and subse-
quently developed as a book (Collins and Evans 2007) and illustrated in a special issue
of Studies in History and Philosophy of Science (Collins 2007), the SEE approach sets out
a normative approach to expertise that starts from the sociological axiom that knowl-
edge is grounded in the life of a community.
This emphasis on experience and socialization serves as the link between SEE
and mainstream work in contemporary science and technology studies (STS), which
we called the Second Wave of STS. Like Wave Two, the sociological approach of
SEE stresses the importance of socialization in the creation, transfer, and applica-
tion of knowledge. This is important because tacit knowledge is an essential part
of any social practice, including science, but the only known way of transferring
tacit knowledge is by social interaction. In other words, without social interaction,
tacit knowledge cannot be acquired and practice will invariably fail in new con-
texts. High-level expertises like interactional expertise therefore depend on mastering
the tacit knowledge needed to speak a language fluently and to respond to novel
problems.
This same emphasis on experience and socialization also marks the SEE perspec-
tive as different from Wave Two STS. Whereas Wave Two is primarily concerned
with documenting how scientific controversies and practice unfold over time, SEE
is more concerned with intervening in real time to make a difference in the way
scientific controversies are understood as they happen. Put slightly differently, SEE
is concerned with the difference between what we have dubbed the problem of
legitimacy and the problem of extension (Collins and Evans 2002). Wave Two STS
directs attention to the problem of legitimacy by showing how the boundary work
of the scientific community and their supporters serves to exclude those with other
kinds of expertise and experience, and shows how the legitimacy of such decisions
could be increased if more heterogeneous forms of participation were developed
(Irwin 1995; Funtowicz and Ravetz 1993). In contrast, SEE is concerned with the
emergent problem of extension, which is created by the inability of Wave Two to
draw a boundary around those who might be counted as legitimate contributors
to technical debate. By classifying and distinguishing between different types of
expertise, SEE aims to rescue the idea of expertise from the corrosive implications
Interactional Expertise and the Imitation Game 55
of STS; Wave Three is intended to show what it means to “know what you are
talking about.”
Three Types of Expertise

Lest this all sound too abstract, imagine that a social scientist wants to investigate the
nature of scientific work in a particular field. Assuming she starts as a novice, the
researcher’s first task will be to understand the ideas, institutions, and actors that make
up the field. She will read books and journal articles and may attend lectures. She will
probably go to conferences and will definitely want to speak to the scientists them-
selves. As the project develops, one marker of the researcher’s growing expertise will
be the quality of these conversations with the scientists. Discussion of the science
being researched will change from simple if rather stilted interviews to increasingly
natural and engaging conversations.
But increasing linguistic fluency need not, and often will not, involve the develop-
ment of practical skills. For example, it is unlikely that the social researcher will ever
take an active role in the substantive research work of the science he or she studies
by designing or running crucial experiments. Similarly, while the social scientist
would be expected to publish in the journals of his or her own discipline, publishing
papers in the field being researched would be unusual and would certainly not be a
requirement.
These differences between novice, social science expert, and practitioner expert
were formalized in the threefold categorization of specialist expertise developed in the
paper that launched SEE (Collins and Evans 2002, 254):
1. No Expertise: That is the degree of expertise with which the fieldworker sets out; it is insuf-
ficient to conduct a sociological analysis or do quasi-participatory fieldwork.
2. Interactional Expertise: This means enough expertise to interact interestingly with participants
and carry out a sociological analysis.
3. Contributory Expertise: This means enough expertise to contribute to the science of the field
being analyzed.
Beginning to categorize expertise in this way leads to two important observations.

First, because expertise is related to experience, and experiences vary, each expertise
must have a distribution. Not everyone can be an expert in everything even if everyone
is an expert in something. Second, because not everyone has the same experiences,
there must be some domains in which at least some individuals have no expertise.
Taken together, these two observations underpin both the problem of legitimacy (i.e.,
the need to recognize expertise where it exists) and the problem of extension (i.e., the
need to recognize the absence of expertise where there is none).
Table 4.1
Periodic Table of Expertises
Ubiquitous Expertises
Dispositions Interactive Ability

Reflective Ability
Specialist Ubiquitous Specialist

Expertises Tacit Knowledge Tacit Knowledge
Beer Mat Popular Primary Interactional Contributory

Knowledge Understanding Source Expertise Expertise
Knowledge
Polimorphic
Mimeomorphic
Meta- External Internal

expertises (Transmuted Expertises) (Nontransmuted Expertises)
Ubiquitous Local Technical Downward Referred

Discrimination Discrimination Connoisseurship Discrimination Expertise
Metacriteria Credentials Experience Track Record
The Periodic Table of Expertises

The three-stage model of no expertise, interactional expertise, and contributory exper-
tise has been expanded into a “Periodic Table of Expertises” (see table 4.1; see also
Collins and Evans 2007, esp. chs. 1 and 2).
The top row of the Periodic Table indicates the ubiquitous expertises that every
member of a society must possess in order to live in it. This is important because the
sociological model of expertise presumes a community that shares a natural language
within which individuals are socialized and where expertise can be generated, held,
and shared. The second row identifies some dispositions or personal qualities that are
necessary for gaining expertise but do not, in themselves, constitute an expertise. For
example, in order to be socialized, one must interact with other members of that
community, but being “easy to get along with” does not make you an expert in any-
thing in particular.
The next three rows are the most interesting. The row labeled “Specialist Expertises”
lists the different kinds of expertise that can be developed in any substantive domain.
Starting at the left-hand side, the table identifies the first and lowest level of specialist
expertise as beer mat knowledge. This corresponds to knowing simple facts that feature
in games like Trivial Pursuit or Who Wants to Be a Millionaire. The name itself derives
from the fact that this kind of information can sometimes be found on the coasters
provided in bars.
The next level of expertise is popular understanding, which refers to the kind of
expertise that can be developed by reading popular accounts of a particular domain.
Unlike beer mat knowledge, in which facts are known but not interconnected, popular
understanding begins to link different facts together. Examples of this genre include
popular science magazines and books that aim to explain science to a general audi-
ence. The next step up is primary source knowledge. The difference here is that knowl-
edge is now based on specialist publications like peer-reviewed journals that are
written by practitioners for practitioners.
Although there is a clear progression from beer mat knowledge to popular under-
standing to primary source knowledge, real specialist expertise requires more. Because
all three competences mentioned so far can be acquired without speaking to a prac-
titioner, they exclude the specialist tacit knowledge held by the expert community of
practitioners. It is for this reason that interactional expertise and contributory expertise,
both of which depend on socialization within the expert community, are qualitatively
different. The difference between primary source knowledge, the highest form under
the left heading (ubiquitous tacit knowledge), and interactional expertise, the first
form under the right heading (specialist tacit knowledge), is vital. The counterintuitive
conclusion that SEE draws from this is that the difference between contributory and
interactional expertise is not great in those many practical settings where experts
interact through words rather than deeds, whereas, in the same circumstances, the
difference between interactional expertise and primary source knowledge is crucial.1
The next row of the table identifies the types of meta-expertise that are a corollary
of the socialization model outlined above. If it is impossible for everyone to be an
expert in everything, then everyday life becomes a puzzle: how do we make choices
about issues in which we have no particular expertise and where experts appear to
disagree? The idea of meta-expertises provides the solution by recognizing that there
are expertises about expertise that can be used in the absence of any specialist exper-
tise. This does not mean such judgments are to be preferred to those of specialist
experts; in many cases they would not be. Nevertheless, meta-expertises are necessary
because without them we would be unable to make many of the choices we have to
make to live in contemporary society.
The different kinds of meta-expertise capture the different ways in which judgments
about experts can be made. As with the types of specialist expertise, there is a distinc-
tion between types of meta-expertise that use ubiquitous tacit knowledge and those
that use tacit knowledge acquired through some specialist training. These differences
can be explained as follows:
• External meta-expertises: These are judgments that are based on the widely shared
ubiquitous tacit knowledge of everyday life. Discrimination refers to the ways in which
experts and the organizations they represent are routinely judged by their look,
demeanor, and reputation. The idea of local discrimination reminds us that some com-
munities will have had additional dealings with particular experts or organizations
before (e.g., because they live near a particular industrial site), and these experiences
will shape their views even in the absence of any significant substantive expertise (e.g.,
previous promises may have been broken). In these cases, even in the absence of any
specialist knowledge, local communities might reach different conclusions from those
reached by more distant communities.
• Internal meta-expertise: This category contains those judgments that require some
appreciation of the criteria used by the experts they judge. Connoisseurship refers to
expertise in consumption rather than production. For example, connoisseurs of wine
would typically be familiar with the conventions and techniques of winemaking and,
therefore, what counts as a good wine, even though they might not be winemakers
themselves. Downward discrimination applies most readily to relatively settled areas of
knowledge and refers to the ability to identify a mistake and, on this basis, to discount
claims as being made by someone with recognizably less expertise. Finally, referred
expertise highlights the ability to use experience acquired in one domain to make
judgments in another. The most detailed example (Collins and Sanders 2007) refers
to the way managers of large, multidisciplinary science projects use their experiences
as bench scientists in other domains to do such things as set realistic targets, judge
between competing claims, and recognize when something is “good enough” to be
an acceptable solution even though other scientists might reject it because it is not
the best possible solution.
The final row of the table identifies some criteria that might be used for identifying
or choosing between experts. The least powerful criteria are qualifications because they
exclude expertise based on experience. Track record is slightly better but also under-
values experience and is unavailable in any genuinely novel setting. Instead, given
the emphasis on understanding expertise as social fluency, the best indicator of exper-
tise is experience, and the more extensive and recent it is, the better.
Interactional Expertise
Each of the categories of expertise set out in the Periodic Table of Expertises could
give rise to a research program in its own right, but, to date, interactional expertise
has been subject to the most systematic research. Interactional expertise is interesting
because it touches on both philosophical issues, such as the relationship between the
body and language, and practical concerns, such as how to do interdisciplinary
research.
The philosophical significance of interactional expertise is the challenge it poses to
the conventional distinction between explicit and embodied knowledge. In the
Periodic Table, explicit knowledge is represented by everything up to primary source
knowledge, while embodied knowledge is found within the category of contributory
expertise. Interactional expertise has no place when explicit knowledge and embodied
knowledge are seen as the only possibilities. For the phenomenologist, for example,
if knowledge is not embodied, it must be explicit.
SEE, on the other hand, implies that there are two different kinds of embodiment.
Traditionally the language and conceptual structure of the world are taken to be a
function of the physical characteristics of the human body (for a discussion of this
point, see Selinger, Dreyfus, and Collins 2007). SEE does not challenge this claim, but
sees it as a matter of the bodily form of the human species rather than the individual.
According to SEE, an individual can gain the interactional expertise pertaining to a
species, including all the relevant tacit knowledge, taken-for-granted assumptions, and
so on, so long as he or she has only a minimal body—i.e., one that can do no more
than hear and speak. In principle, a body can acquire the interactional expertise per-
taining to seeing, feeling, and doing in the way that humans see, feel, and do, without
itself being able to see and feel and do (Collins and Evans 2007, esp. ch. 3; see also
Selinger, Dreyfus, and Collins 2007).
If this were not true, then, as has been argued above, each of us would know only
our private world of sensation. Thus, disabled people would not be able to understand
the world of the able-bodied, and vice versa, and in each case their embodiment would
be immediately apparent from their speech. Still worse, there would be no division of
labor, no interdisciplinarity, and no hope of translation between the languages.
The difference between explicit, interactional, and contributory expertise can be
summed up by reworking the distinction between “talking the talk” and “walking
the walk.” If “talking the talk” corresponds to primary source knowledge (knowing
what has been said), and “walking the walk” corresponds to contributory expertise
(actually being able to perform the task), then interactional expertise corresponds
to “walking the talk”—that is, being able to use the language in novel settings in
much the same way as a contributory expert might. Phrased this way, the defini-
tion of interactional expertise also suggests its own empirical test: Can those who
have it really “walk the talk” or not? It is this ability that is tested by the Imita-
tion Game.
The Imitation Game Method
The Imitation Game originated as a parlor game in Victorian Britain. In this form, the
game was played between a man, a woman, and a judge who could be either male or
female. The man and the woman go to separate rooms and write answers to questions
submitted by the judge. The challenge is for either the man or woman to pretend to
be a member of the opposite gender while the other answers naturally. For example,
the man might have to answer as if he were a woman, while the woman answered as
herself. The task for the judge is to work out, just from the answers to the questions,
who is the real woman and who is the man pretending to be a woman.
The idea of the Imitation Game retains a contemporary relevance because of its use
by Alan Turing as a definition of machine intelligence (Turing 1950). Under the Turing
Test protocol, one of the participants in the Imitation Game is replaced by a computer,
and the challenge for the human judge is to work out which answers are produced
by the real human and which by the computer. Turing claimed that if the computer
succeeded in fooling the human judge for five minutes or more, it should be deemed
intelligent.
There is a clear connection between the Imitation Game and the idea of interac-
tional expertise. Given that the Imitation Game is based solely on sequences of ques-
tions and answers—i.e., it is purely linguistic—a person with interactional expertise
should be indistinguishable from a person with contributory expertise. In contrast, a
person without interactional expertise would not be able to reproduce the discourse
of the contributory expert and so would be readily identified. In other words, success
in the Imitation Game can serve as an operational definition of interactional
expertise.
We have developed a Web-based application that allows Imitation Games to be
conducted over the Internet (details are available from http://www.cardiff.ac.uk/socsi/
expertise). Like most experiments, the Imitation Game is much more complicated in
practice than it first appears. The method is now briefly summarized in a form that is
meant to help those who want to run such games for themselves.
Step One: Decide on the Topic

The first task is to decide on the topic and hypothesis. That is, do you expect that the
person charged with pretending really has interactional expertise or not? If you think
the person pretending has interactional expertise, then the hypothesis is that they
should succeed in the Imitation Game and that the judge will be unable to tell who
is the “real” expert because both are experts (one contributory, one interactional).
Over a series of experiments of this kind, you would expect judges to be right as
often as they are wrong, because they are essentially guessing. This is called a chance
condition run.
If, on the other hand, you think the person doing the pretending is being asked to
display an expertise they do not possess, then your hypothesis will be that the judge
should be able to identify who is who because lack of expertise will be revealed in the
answers of the person who is pretending. This is called an identify condition run.
The two experimental conditions are summarized in figure 4.1. Chance condi-
tions occur when the person claiming to have interactional expertise is (or is
expected to be) well integrated with the domain they are being asked to mimic.
In this case, interactional experts have lots of interaction with contributory experts,
which provides the opportunity for them to learn the language of that community.
In contrast, identify conditions are generated when the person who is pretending
has little or no social contact with the other group. In these cases, the would-be
interactional experts have little or no interaction with the contributory experts and
remain within their own communities. If possible, it is best to run experiments in
which reversing the role of the participants produces a switch from the identify
condition to the chance condition so that the outcome of the two conditions can
be compared.
= Contributory Experts and Judge = Potential Interactional Experts
Chance Condition Identify Condition
Figure 4.1
Chance/identify condition diagram.
Step Two: Recruit Participants

The next task is to recruit participants and perform the experiments. Again, the basic
rules are simple, but the devil is in the detail. In the ideal experiment both the judge
and the person answering naturally will be contributory experts. This is to ensure that
the test is as hard as possible for the person who is pretending. Following this rule,
in the original gender-based Imitation Game, the man pretending to be a woman
should have to fool a female judge rather than a male judge.
It is best if none of the participants knows the identity of the others. That is, even
if the participants are known to each other in real life, they should not know who is
taking part in a specific experiment. The reason for this is to make sure that judges
cannot use information about the particular individuals involved to inform their
guesses; thus, they will be forced to consider the general skills exhibited, such as
familiarity with male or female culture. Using the Internet as the medium of exchange
makes it possible for participants to be located at an arbitrary distance from each other,
and this helps to avoid the problem of judges knowing who is who. On the other
hand, it is vital that the research team can authenticate the identity of the participants,
and some existing reports of false identities being successfully assumed in Internet
chats or games are vitiated by the problem of knowing who is really who.
It can be difficult to recruit participants from small minority groups. In such cases,
identify condition runs (in which both the judge and one other participant has to be
a member of the minority group) are difficult to organize.
Step Three: Data Collection

Once participants have been recruited, the experiment involves a series of questions
and answers. The judge sends the same questions to each participant. When both
participants have typed their answers, these answers appear, simultaneously, on the
judge’s computer screen. The judge then makes a provisional judgment about who
has given which answer and records his or her level of confidence in this judgment.
The judge then sets the next question and the process continues until the judge is
sure they have worked out who is who or feels they cannot make further progress.
Developing the questions is the key to the Imitation Game and raises several inter-
esting issues. For example, while the quasi-experimental nature of the task gives rise
to concerns about demand characteristics and the need for standardization, the logic
of the Imitation Game method leads in the opposite direction to conventional wisdom
in the social sciences. In the case of demand characteristics, for example, the analogy
with a “true” experiment is misleading as there is no direct manipulation of an inde-
pendent variable that demand characteristics might mask.2 Clearly the instructions
given to the participants matter, but the researcher in the Imitation Game can be
completely open about what is required as there is no intervention to hide.
Researchers using the Imitation Game should, therefore, provide judges, contribu-
tory experts, and interactional experts with clear guidance about the role they are to
play. Judges, for example, must know that one person is pretending and that one is
answering naturally. Contributory experts must know they should answer naturally
and (would-be) interactional experts need to know that they must pretend to be
something they are not. For judges, the key requirement is that they know they are
trying to identify a member of a social group rather than a specific individual. In most
cases this can be accomplished by ensuring that participants remain anonymous,
although it may be useful to remind judges that they should not try to work out
individual identities. Interactional experts, in contrast, can be advised that it might
be useful to base their answers on a real person who has the relevant expertise or
experience, if they know one. In this sense, the literature on demand characteristics
(e.g., Orne 1962) supports the experimenter: to the extent that participants want to
“do the right thing,” there is nothing in the Imitation Game method that prevents
the researcher from explaining what “the right thing” is. In other words, the more
perfectly the participants share the researcher’s understanding of what the experiment
consists of, the better the results should be.
Other kinds of advice can also be useful for both the judge and the interactional
expert. For example, questions that require beer mat knowledge can prove problem-
atic. These questions can often help judges discriminate between contributory experts
and those without any expertise, but they do not, as a general rule, reveal much about
the social experience of being a member of a particular culture. In these cases, the
most practical response is to encourage participants to make better use of Internet
search engines and other stocks of explicit knowledge to force judges to ask questions
that touch upon depth of cultural integration rather than mere surface knowledge.
Thus, judges should be reminded that, because the experiment is being run over the
Internet, the person who is pretending can simply look up answers online; thus, a
good strategy is to avoid questions that can be answered in this way. Similarly, the
participants charged with pretending can be reminded that this strategy exists. In this
way, participants can be directed away from questions that rely on explicit knowledge,
however obscure, and toward those that require experience.
The use of the Internet also points to another advantage of letting judges choose
their own questions. While the temptation might be to try to standardize questions,
the role of tacit knowledge in expertise implies that standardization comes with a
significant cost. The reason is that as questions and their answers become standardized
and explicit, they move down the ladder of specialist expertise and so become less
discriminating. Letting judges choose their own questions gets round this problem by
allowing judges to set their questions in the context of their own, up-to-date, lived
experiences. In this sense, the Imitation Game becomes a formalized “ethno-method”
in which participants, be they judges, contributory experts, or interactional experts,
automatically adjust their contributions to reflect the current state of knowledge
within their respective communities.
For all these reasons, the Imitation Game generates a range of data that can be used
to explore similarity and difference across and within cultural groups. At present, the
software developed for the Cardiff experiments automatically records questions,
answers, levels of confidence, and the judge’s reasons for guessing one way rather than
another. More data could also be collected, however. In most cases the experiment
lasts about thirty minutes, but this includes several fairly lengthy waits. For this reason
it is often useful to have researchers with the participants who can use these pauses
to collect more detailed data about the strategies being used by the judge or the expe-
riences participants are drawing on in order to compose their answers. Alternatively,
the opportunity could be taken to record basic face-sheet data and other details that
might help with interpretation and the contextualization of individual results.
Step Four: Data Analysis

When the series of experiments is complete, both quantitative and qualitative data
are available. The quantitative data consists of the guesses and levels of confidence,
and can be used to see if the hypothesized differences between the chance and identify
conditions have emerged in practice. In the ideal scenario, the identify condition
would show 100 percent correct guesses, all with very high confidences. In contrast,
results from an ideal chance condition would show equal numbers of right and wrong
guesses, with the proportion of “Don’t know” responses reflecting participants’ will-
ingness to make definite judgments. Because of the variation in “Don’t know”
responses, the fundamental measure of successful identifications is not the absolute
number of correct guesses, which will be influenced by the propensity of participants
to make confident guesses, but the excess of right guesses over wrong guesses (i.e.,
“confident and correct” minus “confident and incorrect”) as a proportion of the total
number of Imitation Games in that condition. This number, which we call the “iden-
tification ratio” (IR), can then be used to compare results across topics and over time.
In practice, the differences are never as stark as those between the ideal types, but it
is often quite easy to see that the two conditions are not the same and, so long as the
sample size is large enough, these differences are generally statistically significant.
The qualitative data consists of the sequences of questions, answers, and judgments
contained in the text captured by the software plus the field notes and other records
made as participants were interviewed and observed. This qualitative data can be used
to answer a number of different questions. For example, looking at the topics used in
the questions provides an insight into what judges think differentiate the two social
groups, i.e., what it is that makes their group unique and what kind of insights will
be beyond the capacity of the person who is pretending. In contrast, looking at the
answers shows what kinds of knowledge are shared and how people with interactional
expertise are able to reproduce the discourse associated with experiences they have
never had. When complemented by interview and field note data, these sequences of
questions and answers provide powerful insights into the diversity of different cultures
as well as how they overlap and intermingle.
Even though its primary purpose is to investigate interactional expertise, the Imita-
tion Game also allows exploration of some other categories in the Periodic Table of
Expertises. For example, looking at the chance condition runs and the reasons judges
give for choosing between apparently similar responses provides some insight into the
strengths and limits of discrimination. This is because some chance condition runs
do produce correct identifications and some produce very confident guesses. Examin-
ing how these guesses are made provides insights into folk theories of lying (e.g.,
Which are an indication of honesty: short answers or long answers?) and the ways in
which minor discursive cues are built up into strong convictions.
Boosting the Sample Size: Phase Two Experiments

Running Imitation Games in real time can be logistically complex. For the ideal experi-
ment you need three participants and three researchers to be available at the same
time. Fortunately there is a way to gather additional data from the initial live runs:
the transcripts from real-time runs can be collected and sent to new judges, so that
new judgments can be made which will confirm or disconfirm the live data. Using
this approach, much larger data sets can be created and a larger number of judges
recruited than if only real-time experiments are used, boosting statistical significance
as well as subjective confidence in the validity of the outcomes. In addition, as the
sample size increases, the sophistication of statistical analysis can increase as it becomes
possible to distinguish between different subgroups.
In a case where very few live runs have been conducted and these include system-
atic errors, there is a danger that a spurious result can be given unwarranted statistical
significance because “secondary” judges, reading only the flawed transcripts, will be
in a position to make only the same judgments as the original judges. Secondary judges
could thus reinforce a “chance” judgment by being unable to make good guesses from
the transcript of a poorly conducted initial interview. Likewise, they could reinforce
an “identify” judgment by making the same correct guess which was the obvious
outcome of a poor answer given by a respondent. In such cases, the problem of
secondary judgments is akin to what psychologists call a “stacking effect.”
Empirical Examples
So far the Imitation Game has been used by the Cardiff group to investigate a number
of minority groups with a view to demonstrating the idea of interactional expertise,
testing for the possession of interactional expertise, and testing for membership of
social groups. The best known of these is probably the Imitation Game involving Harry
Collins and gravitational wave physicists, which has been featured in Nature and
several Internet news sites (see Giles 2006; slate.com). In the experiment, Collins, who
has been researching the gravitational wave physics community for almost three
decades, and a gravitational wave physicist provided answers to questions posed by
another gravitational wave physicist. The two sets of answers were then shown to a
series of other gravitational wave physicists, who were asked if they could tell which
answers came from Collins and which from the physicist. The result was that the
physicists were unable to reliably identify Collins, demonstrating both that interac-
tional expertise exists and that intensive social science fieldwork is one way in which
it can be acquired.
The other experiments that have been written up and published, this time in the
peer-reviewed literature, are on color blindness and perfect pitch. The results have
been described in detail elsewhere (Collins et al. 2006; Collins and Evans 2007), so
what follows here is only a brief summary. Color blindness was chosen because the
ability to distinguish between colors is taken for granted by the majority of the popu-
lation, and everyday discourse reflects this. There are, however, a small number of
people (about 5 percent of the population) who are unable to distinguish between
some or all colors and who are, therefore, classed as color-blind. The hypothesis is
that, because the color-blind have been brought up and socialized within the society
of color perceivers, their language will be indistinguishable from those with normal
color vision. In the Imitation Game, the color-blind should succeed in fooling a judge
because they will have interactional expertise in color language. If the roles were to
be reversed, and a person with normal color vision were asked to pretend to be color-
blind, we would expect them to fail. In this case, the hypothesis is that, because most
color perceivers have little or no experience of the problems faced by the color-blind,
they will lack interactional expertise.
In the case of perfect pitch, the predictions are inverted. Perfect pitch is the ability
to identify the musical note that corresponds to a sound, but, unlike normal color
vision, perfect pitch is relatively rare, and so the reduced form of perception (“pitch-
blindness”) is the norm. In the Imitation Game we would thus expect people with
perfect pitch to be able to mimic the absence of this ability. In contrast, if the roles
were to be reversed and people without perfect pitch—the “pitch-blind”—were asked
to mimic the discourse of perfect pitch, they would fail for the same reasons that
normal color perceivers cannot pass as color-blind.
Using these two groups created four different experimental conditions—two chance
conditions and two identify conditions—based on whether or not the experience of
the judge was expected to be shared by both of the other participants (see table 4.2).
The chance conditions represent the “proof of concept” runs, for the hypothesis is
that the pretender has the interactional expertise needed to successfully imitate having
the target experience. The identify conditions act as a kind of experimental control
by demonstrating that, where the socialization is absent, so too is the expertise.
As detailed elsewhere, the results of these experiments were consistent with the
hypothesis, and statistically significant differences were found between the chance
and the identify conditions. Our conclusion is, therefore, that the results support the
claim that interactional expertise exists: the color-blind can pass as color perceivers
relatively easily because they have been immersed all their lives in the language of
color. In contrast, the pitch-blind cannot pass as pitch perceivers because they have
not been so immersed.
In more recent work, which we hope to publish in due course, we have run Imita-
tion Games on a number of other topics. These have included the extent to which
the blind can pass as sighted, and vice versa; whether members of ethnic minority
communities can pass as white, and vice versa; and whether or not gay and lesbian
participants can pass as heterosexual, and vice versa. In addition, Theresa Schilhab, at
Table 4.2
Expected Outcomes of Imitation Game Experiments
Pretender is Target Expertise Expected Outcome
A: Color-blind Imitates Color-perceiving Chance

B: Color perceiver Imitates Color-blind Identify
C: Pitch perceiver Imitates Pitch-blind Chance
D: Pitch-blind Imitates Pitch-perceiving Identify
Adapted from Collins and Evans 2007, 96.

the Danish School of Education, has run Imitation Game experiments on whether
midwives who have not given birth can pass either as lay mothers or midwives who
have given birth. Taken together, this research suggests a bright future for the idea of
interactional expertise and the Imitation Game method.
Conclusions: Interactional Expertise in Action
The Imitation Game experiments bear in two ways on the concept of interactional
expertise and its relationship to the wider typology of trading zones set out in chapter
2. First, they demonstrate that “walking the talk” is indeed possible and that interac-
tional expertise is an observable empirical phenomenon. The Imitation Games provide
positive support for the idea of trading zones in general and for the development of
fractionated trading zones in particular.
The experiments also showed that nothing can be taken for granted. For every
chance condition we created, there was a corresponding identify condition character-
ized by the absence of a shared language to carry disparate experiences across social
boundaries. While there is nothing in principle that prevents the interactional exper-
tise needed for one group to understand the other from developing, acquiring this
expertise is difficult. In the case of the chance conditions we have studied, the partici-
pants with interactional expertise had all been immersed in the majority/contributory
culture for periods measured in years, not months or days. Transferred into the context
of interdisciplinary research, multiagency teams, or a multicultural society more gener-
ally, the clear implication is that those charged with creating mutual understanding
must be alert to the scale of the problem and not underestimate the time and effort
needed to understand the perspective of another. In the case of research funding, in
particular, where interdisciplinarity is often encouraged, research projects will need to
include specific time periods for gaining the interactional expertise needed to under-
stand each discipline’s concerns and problems.
By demonstrating that interactional expertise is real, the Imitation Game experi-
ments help us understand how many of our existing social institutions work. Peer
review groups, advisory committees, and multidisciplinary teams of all sorts already
function because their members are able to communicate with each other. What the
idea of interactional expertise provides is an explanation for why these teams and
groups can function that does not founder on either the absence of tacit knowledge
that undermines any explanation based on formal knowledge alone or the need for
actual experience that the phenomenological approach claims to be a necessity. By
being both laden with tacit knowledge and only minimally embodied, the idea of
interactional expertise provides both the glue that binds social groups together and
the lubricant that allows them to mix and mingle.
Notes
1. The distinction between polimorphic and mimeomorphic actions, which appear beneath the
row of types of specialist expertise, refers to the extent to which actions—in this case, expertise—
can be mimicked by machines. Actions that can be reproduced by machines are said to be
mimeomorphic. Actions that rely on an understanding of tacit social rules and cannot, therefore,
be reproduced by machines are said to be polimorphic. The distinction is not particularly impor-
tant in this context and is explained in more detail in Collins and Kusch 1998.
2. To the extent that there is such a variable, it is the socialization of participants prior to taking
part in the experiments.
References
Collins, Harry. 2004. Interactional Expertise as a Third Kind of Knowledge. Phenomenology and
the Cognitive Sciences 3 (2):125–143.
Collins, Harry, ed. 2007. Case Studies of Expertise and Experience. Special issue of Studies in
History and Philosophy of Science 38 (4).
Collins, Harry, and Robert Evans. 2002. The Third Wave of Science Studies: Studies of Expertise
Collins, Harry, and Robert Evans. 2007. Rethinking Expertise. Chicago: University of Chicago Press.
Collins, Harry, Robert Evans, Rodrigo Ribeiro, and Martin Hall. 2006. Experiments with Interac-
tional Expertise. Studies in History and Philosophy of Science 37 (4):656–674.
Collins, Harry, and Martin Kusch. 1998. The Shape of Actions: What Humans and Machines Can
Do. Cambridge, MA: MIT Press.
Collins, Harry, and Gary Sanders. 2007. They Give You the Keys and Say “Drive It!” Managers,
Referred Expertise, and Other Expertises. Studies in History and Philosophy of Science 38
(4):621–641.
Funtowicz, Silvio O., and Jerome R. Ravetz. 1993. Science in the Post-Normal Age. Futures 25
(7):739–755.
Giles, Jim. 2006. Sociologist Fools Physics Judges. Nature 442 (8).
Irwin, Alan. 1995. Citizen Science: A Study of People, Expertise and Sustainable Development. London:
Routledge.
Orne, Martin T. 1962. On the Social Psychology of the Psychological Experiment: With Particular
Reference to Demand Characteristics and Their Implications. American Psychologist 17:776–783.
Ribeiro, Rodrigo. 2007a. Knowledge Transfer. Ph.D. dissertation, Cardiff University.
Ribeiro, Rodrigo. 2007b. The Language Barrier as an Aid to Communication. Social Studies of
Science 37 (4):561–584.
Selinger, Evan, Hubert Dreyfus, and Harry Collins. 2007. Interactional Expertise and Embodi-
ment. Studies in History and Philosophy of Science 38 (4):722–740.
form. Studies in History and Philosophy of Science 38 (4):642–656.
Turing, A. M. 1950. Computing Machinery and Intelligence. Mind 49:433–460.

II Applying Trading Zones and Interactional Expertise to Domains
of Practice
Applying the framework of trading zones and interactional expertise to application

areas hones and refines the framework while also determining its practical value.
Chapters in this part of the volume include applications to service science, business,
the environment, and education.
Michael E. Gorman and Jim Spohrer’s chapter on service science looks at one pos-
sible new field that might emerge out of an astonishingly wide range of existing dis-
ciplines, focused on the future of services. The motivating problem is how the producers
of services and their clients can create together new sociotechnical solutions that solve
existing problems or create new opportunities. This approach changes the old model,
in which clients demand and producers provide, into a collaboration between produc-
ers and clients, where the roles themselves begin to blur. IBM and other organizations
are promoting a new field of service science, management, and engineering (SSME).
The central question of Gorman and Spohrer’s chapter is whether such a new field is
necessary, and what SSME expertise might encompass.
Jeff Shrager discusses the ways software has increasingly become a mediator bridg-
ing the chasm between scientists and reality. In his chapter, Shrager outlines three
decades of attempts to bridge a related chasm—between scientists and their instru-
ments. In order to use modern scientific instruments, scientists often have to work
with programmers.
In order to bridge this chasm for one field, Shrager decided to go beyond the acqui-
sition of interactional expertise by actually becoming a molecular biologist to see if
dual training was the answer: perhaps molecular biology and computer science should
be combined. He made a deal to do computational work for biologists in exchange
for being trained to do their sort of research—a very simple trading zone. Shrager
succeeded in learning enough molecular biology to become a contributory expert, but
he found most of his time was spent doing computational work because the biologists
did not become familiar with his expertise. At this point he became aware of the extent
72 Applying Trading Zones and Interactional Expertise to Domains of Practice
of the chasm between biology and computation. His next attempt was to design
BioBike, a tool that biologists could use easily to perform their own computations.
Though BioBike has been very useful to Shrager and other computer scientists helping
biologists, it has not to date been picked up by the biologists themselves as a stand-
alone tool. BioBike has, however, facilitated trading zones between biologists and
programmers, acting as a kind of hybrid of a creole and a boundary object.
Brad Allenby’s chapter focuses on the anthropogenic earth, arguing that what has
traditionally been regarded as nature is part of a global sociotechnical system. Allenby
pays particular attention to the way in which ideology can prevent a trading zone—
not everyone wants to trade—and he himself criticizes the idea of the trading zone
because it depends on “enlightenment and Western values.” The activity of trading
in areas where cultures make contact is as old as civilization, but Allenby reminds us
to be aware of the assumptions behind the current use of the term, especially when
applied normatively.
Lekelia D. Jenkins’s chapter applies the framework of trading zones and interac-
tional expertise to the problem of reducing marine bycatch. For example, the kinds
of nets traditionally used for catching shrimp also accidentally catch sea turtles, which
are considered bycatch in this situation. Jenkins uses the trading zone and interac-
tional expertise framework to diagram the trajectory of attempts to reduce turtle
bycatch, starting with dominance by the National Marine Fisheries Service but pro-
gressing to a more equal trading zone, in which the experience of the fishers was
incorporated into at least one device design. There was even a boundary organization
that acted as a kind of interactional expert (the Turtle Excluder Device Voluntary Use
Committee). Jenkins proposes that a new trading zone had to develop to ensure adop-
tion of the new Turtle Excluder Device design; yet, even though agreement was
reached among representatives of key stakeholders, this agreement was repudiated by
the broader community. Jenkins concludes with a new diagrammatic representation
of types of trading zones and the trajectories among them, thereby using her case
study to revise the framework in ways that will drive future research.
Matthew M. Mehalik applies the trading zone framework to a project designed to
produce a scorecard, or set of metrics, for a large urban public school district in a city
in the western United States. He connects trading zones to actor network theory,
focusing on two types of trading zones: those where the participants trade from a
position where they do not share a common mental model of goals and outcomes,
and those where they do share such mental models. Like Jenkins, he uses a graphic
representation of the changes in a trading zone; where Jenkins used graphs, Mehalik
uses network diagrams. One unique feature of Mehalik’s study is that he shared his
Applying Trading Zones and Interactional Expertise to Domains of Practice 73
network diagrams with participants in the trading zone, and encouraged them to
revise them. This process facilitated the identification and elimination of resource
bottlenecks; it also provided individuals with opportunities to generate meaning in
their participation in district reform measures, without requiring them to adopt rigidly
imposed policies.
Erik Fisher and Roop L. Mahajan describe the formation of a fractionated trading
zone around the idea of humanistic engineering, which began with a boundary object
in the form of a gallery exhibit linking “art, engineering and society.” Fisher, the
humanities advisor of the College of Engineering and Applied Science, and Mahajan,
the newly appointed dean of the college, met to discuss this exhibit, and out of that
a collaboration was born. Fisher was embedded in Mahajan’s thermal and nanotech-
nology research laboratory: his desk was located in the laboratory; he attended meet-
ings, participated in equipment training, and met regularly with researchers. Fisher
gradually acquired interactional expertise in nanoscale fabrication and characteriza-
tion, as evidenced by the fact that outsiders often assumed he was an engineer. He
and Mahajan and the researchers traded around Mahajan’s idea of a research protocol
that would seamlessly link humanities and engineering, moving from a potentially
top-down coercive trading zone to a fractionated one in which the parties were rela-
tively equal. Fisher concluded that this fractionated trading zone would not lead to
a creole or a research protocol or a new field. Instead, it promoted reflection on the
part of both the humanist and the nanoscientists, resulting in improved research
results for both. A fractionated, collaborative trading zone can be a worthwhile end
in itself.
Bolko von Oetinger takes the trading zone framework and applies it to one of the
central problems in business strategy: How can successful firms learn to cross paradig-
matic boundaries? Typically, a disruptive change in the business climate is signaled
by some group or interest on the periphery of the successful firm’s business and intel-
lectual territory. Trading zones between well-established firms and more innovative,
peripheral players are one way to get ideas across traditional boundaries. The zone
serves as a kind of no-man’s-land where the ordinary norms of competition can be
suspended. Von Oetinger uses examples like the rise of Internet music, which was at
first resisted by the major labels, which focused on selling CDs as they had earlier sold
albums. Here Steve Jobs may have served as a kind of interactional expert, convincing
the major labels that they could make money by selling music over the Internet.
Trading zones are necessary in business only when all parties stand to lose or
gain together. In situations where there can be winners and losers, imitation is often
a better strategy than trading, as exemplified by reverse engineering. But there are
74 Applying Trading Zones and Interactional Expertise to Domains of Practice
situations, such as the current financial crisis, in which trading zones could be most
helpful, especially ones that include not only business but government and other
stakeholders. Von Oetinger concludes with a provocative set of questions about
boundary crossings that could serve as a basis for a trading zone between science-
technology studies and business administration.
5 Service Science: A New Expertise for Managing Sociotechnical
Systems
Michael E. Gorman and Jim Spohrer
Introduction
There are three main characteristics of services (Davis and Berdrow 2008):
1. Services are actions rather than goods;

2. Customers are involved in the production of value;
3. When the customer relationship ends, so does the value, unless lessons are captured
from the experience for the next service exchange.
Service systems, including government, health care, education, retail, and professional
consulting, are the fastest growing sector of the global economy—especially informa-
tion and business services (Spohrer et al. 2007). The service component of the major
industrialized economies is greater than 50 percent and growing, and developing
nations are close behind (Paulson 2006). Services comprised 80 percent of U.S. eco-
nomic activity as of 2003 (National Academy of Engineering 2003, 8). What is true of
the overall economy is also true of the businesses that make up the economy. Com-
panies such as GE and IBM, which initially focused on manufacturing, now realize
more than 50 percent of their revenues from service activities. This trend illustrates
the growing market for figuring out what a subset of clients need and creating high-
value, custom solutions. Governments, businesses, and individuals all need help
managing complexity in a time of accelerating technological change (IfM and IBM
2007). These conditions create the need for a new kind of expertise in service sciences,
management, and engineering (SSME) (Chesbrough 2004; Chesbrough and Spohrer
2006).
A good working definition of “service” is the co-creation of value via client-provider
interactions. Note the emphasis on the collaboration between provider and client.
Service systems are sociotechnical networks: human beings, technologies, and organi-
zations are closely coupled (Elzen and Enserink 1996). A service actor, whether client
76 Michael E. Gorman and Jim Spohrer
or provider, may be an individual (person), machine (owned by a person or business),

business, organization, or other service system.
The service actor is therefore part of a dynamic network of both human and non-
human participants. Human actors differ from nonhuman participants in their ability
to reflect and make choices (North 2005). If the object of the service is a person or
organization, the service may achieve an immediate goal of that actor or ensure future
achievement (“a promise”), instill a new goal in the actor, transform a capability of
the actor, or instill a new capability in the actor. Thus, service interactions typically
transform some state or attribute of the client actor. The provider may also be trans-
formed by the process.
Mutual benefit or value is co-created via relationships that share work, risk, infor-
mation, assets, and/or decision making. Often, service providers have unique or spe-
cialized knowledge that allows them to perform the service better, faster, and cheaper
than the client could possibly do it on their own. As demand for such “asymmetric
knowledge” services grow, businesses may embody the knowledge in special tools or
organizations to achieve scale efficiencies. Ultimately, if the tools become easy enough
to use, the provider may make the tools available to the clients, resulting in eService
or self-service systems.
Consider, for example, the way in which the Lego toy company lets its customers
propose new designs and share them for modification by other users, creating a kind
of open source design environment (http://factory.lego.com/). Lego engineers then
implement the designs. Here customers provide the designs, within familiar
constraints—the customers have built Lego models, and thus know what is and
isn’t possible. The coevolution is in the collaboration between users and designers,
and the IT needed to facilitate it.
After playing a key role in founding computer science, IBM now believes that there
is a need for a new kind of expertise in SSME, where the engineering component would
include design as well. Obviously, such a broad effort requires a multidisciplinary
approach that spans academic silos. Interdisciplinary efforts can lead to generalists
who eventually become specialists in a new field (Spohrer, McDavid, Maglio, and
Cortada 2006). Something like this happened in computer science, which combined
software and algorithm complexity theory, as well as hardware and logic design, into
a new specialty that increases our understanding of computation in technological
systems. Participants in a recent symposium (IfM and IBM 2007) from multiple disci-
plines and organizations concluded that “Service science has the potential to be as
important as the foundation provided by physics, chemistry, biology, cognitive science
and computer science for the modern world.”1 Other companies are picking up on
A New Expertise for Managing Sociotechnical Systems 77
this initiative, as are multiple universities, primarily in the United States but also in
Germany, Japan, India, and China.2 The U.S. National Science Foundation also has a
program in Service Enterprise Engineering.3
The deeper understanding of service system evolution could lead to more system-
atic approaches to service innovation. Service innovations have the potential to
impact service productivity, service quality, and rates of growth and return for service
systems.
The Service Scientist as an Interactional Expert
IBM, Oracle,4 and other companies believe that the students being produced by uni-
versities today are not prepared to become service scientists, suggesting that existing
domains of knowledge and practice are inadequate to do SSME. New legislation in the
United States has called for promotion of this field, although the expertise is defined
through “a laundry list” of different fields that would be involved.5 How might a truly
novel SSME expertise emerge? How would we know it if we saw it?
SSME may eventually become an expert specialization in its own right, but at the
present time it draws on multiple expertise communities and spans conventional
disciplines. Furthermore, even if SSME becomes a specialized community, its practi-
tioners will have to work with a wide variety of stakeholders from other communities.
Therefore, anyone who practices SSME will have to acquire the ability to share knowl-
edge and practices across disciplines.
Collins and Evans have labeled this kind of ability “interactional expertise,” and
placed it on a continuum between having no expertise in a domain and having con-
tributory expertise (Collins and Evans 2002).6 Let us use Collins’s study of gravitational
wave physics as an example (Collins 2004a). Collins has contributory expertise in the
sociology of science. When he began his work with gravitational wave physicists, he
possessed no expertise in this area. In order to understand their work and their com-
munity, he had to gain interactional expertise.7 In the end, he could not do the
experimental or mathematical work of the contributory experts, but he could carry
on extended, provocative conversations with them and even answer verbal questions
about the field as fluently as a contributory expert (Collins et al. 2006).
Consider another example. Lekelia Jenkins has studied the ways in which some
scientists and fishers have to gain interactional expertise in each other’s communities
in order to design devices that will prevent unwanted species from being caught in
fishing operations—e.g., dolphins, which are often accidentally trapped by purse seine
fishers (Jenkins 2007). A device designed by a scientist in a laboratory is often poorly
suited to deployment on a fisher’s boat. Devices designed by fishers work better on

the boats, but can be improved by interaction with engineers and scientists. Dave
Kreutz, a banker turned fishing boat skipper, has gained interactional expertise with
the scientific and regulatory communities and is able to impact both research into,
and implementation of, bycatch reduction devices.8
Kreutz’s experience could be turned into the kind of case study that illustrates how
an interactional expert emerges to satisfy a growing service need, in this case to
prevent the wrong species from being accidentally killed by fishers. Development of
interactional expertise is facilitated by exposure to case studies from the service’s
frontier that illustrates the ways in which technical and social issues become inter-
twined.9 These case studies can begin as research projects that help identify where
interactional expertise will most likely be needed and what core disciplines are
involved. As the number of different disciplines required for state-of-the-art service
innovation expands, so will the need for different kinds of interactional expertise. So
perhaps Dave Kreutz is a model for SSME—someone trained as a banker and a skipper
who could also talk the languages of the regulators—and able to act as a negotiator
among the communities, with credibility in each.10 SSME should develop methods
and best practices for encouraging the development of this sort of interactional exper-
tise in services.
In order to co-create value, the service scientist is going to have to acquire the
ability to interact not only with the client but also with the client’s domain of exper-
tise. But unlike the sociologist’s situation, in service science the relationship should
ideally be symmetric—the client should also learn to interact with the service scien-
tist’s expertise. Both client and producer may have to acquire interactional expertise
in each other’s domains so they can work together to evolve new solutions not previ-
ously imagined by either. Thus, there will be service science relationships that trans-
form existing knowledge.
Therefore, interactional expertise will be a core capability that needs to be possessed
to some degree by virtually all SSME practitioners. Interactional expertise is similar to
T-shaped expertise, where the tall “leg” of the T is depth in one area of expertise and
the shorter crossbar on top indicates the ability to interact at some level of depth with
those from other expertise communities.11 “SSME qualifications, which we see as criti-
cal to developing adaptive innovators with a service mindset and service innovation
skills, should include interactional skills across the main disciplines of service science.
Interactional skills enable proficiency in the concepts and vocabulary for framing
problems and discussing potential solutions across disciplines.”12
If SSME is primarily an interactional expertise, that has profound implications for

our educational system, which tends to be good at producing disciplinary expertise,
but not interactional skills.13 It will be difficult for many individual practitioners to
gain interactional expertise in all of the disciplines of SSME, which include (at the
very least) service economics, computing, engineering, and marketing, not to mention
relevant portions of science and technology studies (STS). Gaining interactional exper-
tise is a journey that lies on a continuum. One course in each of the key disciplines,
designed to provide the top concepts in each along with real-world discussion cases,
could be practical, with a single year of education added onto an existing home
discipline.
Becoming an interactional expert in service science would also require a participant-
observation experience in at least one disciplinary community outside of the student’s
major. Such experience could be acquired during a summer internship or a semester
abroad, or even by going into a laboratory or incubator during the academic year. In
order to be effective, this experience would have to be accompanied by proper men-
toring on how to gain interactional expertise (Collins 2004b).
A smaller step might be to take a course or two along with a field experience that
illustrate how disciplines and expertises have to be integrated to coevolve value. The
goal of such a program would be to motivate students to acquire the necessary inter-
actional expertise over a lifetime of continuous learning.
A larger step might be possible in even less time, if a fully formed theory existed
to lay the foundation for more easily integrating and learning the range of disciplines.
As we uncover the “real” contributory expertise of service science—a deep understand-
ing of service systems, value propositions, and governance mechanisms—we hypoth-
esize that teaching interactional expertise in the other areas will be greatly simplified.
The end result would be the development of collaborative innovators, capable of
evolving new solutions with clients who are driven by a need.
Trading Zones
According to Kuhn (1970), scientists operating from different research paradigms even
within the same discipline have difficulty communicating because they have con-
structed different versions of reality. Because they operate from different assumptions,
communication between them can only be partial. These “incommensurable” para-
digms (to use Kuhn’s term)14 are exacerbated when one crosses the kind of disciplinary
and expertise boundaries that service science will have to span.
When Peter Galison studied the development of radar and particle accelerators, he
found that different expertise communities had to develop trading zones. The key was
developing a common language to get around barriers to collaboration—first a jargon,
then a pidgin, and finally a full-scale creole.15 Galison argues that it is possible for
communication to take place locally even when participants in a trading zone come
from communities that operate in different paradigms: “They can come to a consensus
about the procedure of exchange, about the mechanisms to determine when goods
are ‘equal’ to one another. They can even both understand that the continuation of
the exchange is a prerequisite to the survival of the larger community of which they
are part” (Galison 1997).
The service scientist will have to facilitate such trading zones between different
disciplines and stakeholders to coevolve solutions to emerging problems and oppor-
tunities. One interesting question is whether SSME will evolve its own paradigm,
including the kind of textbook reconstructions of the field that serve as exemplars for
students, indoctrinating them into proper practices. If SSME becomes a paradigmatic
expertise, paradoxically, trading zones will be even more essential for the success of
the new field—because SSME depends on the coevolution of new solutions with clients
who will likely come from separate and initially incommensurable paradigms.
SSME experts will have to develop a metacognitive awareness of their evolving
paradigm so that they can envision alternatives when confronted by a client with a
radically different view. This capability corresponds to what Johnson (1993) calls
“moral imagination.” The models on which we base much of our ethical reasoning
are often implicit and can become confused with reality for those who hold them.
Moral imagination begins with the recognition that these realities, these truths, are
views. Once I know that I have a view, I can listen to yours without immediately
dismissing it, and we now have the potential to learn from one another. But if I think
I see reality and that you only have a view, then there is no possibility of deep com-
munication. Applied to information technology, moral imagination would require
service scientists in a trading zone to set aside their own firm notions of what the
client needs and be able to listen to alternate views. Here moral imagination converges
with successful service—listen to the client, and try to figure out what he really needs,
not just what he wants.
A Small Nanotechnology Trading Zone
With a small amount of funding from the NSF, Gorman created a trading zone moti-
vated by using nanotechnology to make the world a better place. He shared an engi-
neering graduate student with a materials scientist, and their thinking processes were
documented by Jeff Shrager, a cognitive scientist (Gorman, Groves, and Catalano
2004). In order to trade, members of this small zone had to develop a shared jargon.
Instead of creating new terms, they shared meanings for existing ones like “guided
self-assembly” from the engineering side and “moral imagination” from the ethics
side; the former refers to ways in which molecules can be guided to assemble them-
selves in a pattern on a surface, and the latter will be discussed below. Gorman’s
understanding of a term like “guided self-assembly” was never as deep or experiential
as Groves’s, and vice versa with respect to the concept of moral imagination. Yet each
acquired some interactional expertise in the other’s area.
The team also had to develop a metaphoric language to talk about its goals. All
three participants in this zone liked hiking, and so this activity provided what
seemed to all to be a natural set of metaphors. Groves took the lead in creating
the language. “Distant mountains” referred to major global problems and opportu-
nities, such as human health, climate change, the prevalence of warfare, and so on.
“Closer foothills” represented specific aspects of these problems, like the elimination
of heart disease or cancer, or providing more data on toxins introduced into the
environment as either pollution or a form of biological warfare. The team wanted
to build a bridge that they or others could use to reach a range of local mountains,
or foothills. This bridge would be part of a trail, but could also give access to other
trails.
This set of metaphors grew as the team added collaborators to the trading zone.
Patricia H. Werhane, an expert on business and engineering ethics, suggested think-
ing about an alternate goal to the mountains—perhaps a village in the developing
world, which would direct the project more toward Amartya Sen’s emphasis on
increasing women’s affluence and education worldwide (Sen 1999, 1994). The research
topic selected by the first student incorporated nanotechnology, biotechnology,
and cognitive reflection—three of the four nano-bio-information-cognitive (NBIC)
technologies—and included an explicit ethical component (Gorman, Groves, and
Catalano 2004).
Metaphors may often complement creoles in trading zones that span a particularly
wide disciplinary range. In NSF workshops to develop new interdisciplinary initiatives,
“One of the most striking features of the workshop process is the amount of reciprocal
adjustment required to get all participants, from within NSF and without, talking
about the same topics in a mutually comprehensible language” (Hackett 2000, 254).
These workshops are trading zones, where the participants are funded to work together
but need to adopt at least a shared jargon, as well as to develop metaphors that can
“help groups of people from disparate backgrounds think about a problem in the same
way” (Palmer 2001, 12).
For example, in a workshop Gorman conducted on scientific and technological
thinking (Gorman, Kincannon, and Mehalik 2001), spanning disciplines such as psy-
chology, cognitive science, philosophy, history, and sociology, the pidgin consisted
of agreeing on meaning for certain terms like “problem space” and “mental model.”
The workshop also adopted two primary metaphors: shared toothbrushes and spheri-
cal horses. The former, described by Chris Schunn, referred to the perception that
most scholars liked to share frameworks about as much as they liked to share tooth-
brushes. The latter, described by David Gooding, referred to a competition between a
physiologist, a geneticist, and a theoretical physicist to see who could predict the
outcome of a horse race. After a year, the first two admitted there were too many
variables, but the physicist produced a prediction accurate to seven decimal points—
valid for a perfectly spherical horse moving through a vacuum. These metaphors
emerged early in the workshop and kept the participants aware of the importance of
developing a shared framework (not a toothbrush) and conducting research that was
based on a deep understanding of science and engineering practice (not just of spheri-
cal horses). By the end of the meeting, all a participant had to say was “shared tooth-
brush” or “spherical horse” and everyone else in the room knew what was implied.
More empirical work needs to be done on the relationship between metaphors and
interlanguages in trading zones.
The examples in this part show that it is possible to form interdisciplinary trading
zones on the cutting edge of science and technology. In order to succeed, service
scientists, managers, and engineers will have to facilitate such trading zones,
reflexively studying how to improve them as they create them.
Boundary Objects
Another way to achieve partial coordination in a trading zone is via the use of bound-
ary objects. Consider a simple exchange of a vital resource: “water has a vast range of
meanings, uses, and values—in the abstract or ethical sense of the term ‘values.’ For-
tunately, two strangers who exchange water do not need to engage with the meanings,
uses, and values each imputes to the good they are exchanging. All they need to do
is agree on a price: how much water against how much of something else. In this lies
the simplicity of market exchange: what matters is the scarcity of water for each party,
relative to the scarcity of anything else they want” (Seabright 2004, 128). Water, in
this case, serves as a boundary object for the parties in the exchange.
Boundary objects can be represented differently by multiple participants in a

trading zone and yet have enough of a common representation “to maintain a
common identity across sites” (Star and Griesemer 1989). In the case of currency, there
is a whole set of shared representations about assigning exchange value to coins or
items, but each of the participants sees local features of its currency that the other
party cannot. The participants “remain just as far apart (culturally) at the end of the
interaction as they were at the beginning, but each party manages to use the boundary
object to their own advantage” (Ribeiro 2007).
Parties can also gradually come closer in their representation of a boundary object.
Consider a trading zone in which JPL engineers and scientists tried to reach agree-
ment on where to land a Mars polar probe. The team making the decision included
an internal JPL customer and a consultant who advocated opposite solutions, the
customer favoring the north pole on scientific grounds, the consultant favoring the
south pole due to its engineering feasibility. The Polar Lander served as a boundary
object for the participants in this trading zone, because its orbital trajectory and
construction would be different depending on which pole it was sent to. The par-
ticipants conducted a series of trades around this boundary object and eventually
reached agreement.16
SSME expertise is not necessary for simple exchanges involving tangible objects or
products like water; it is most important when existing options are inadequate and
new trading zones have to be formed to create solutions.17
A Taxonomy of Trading Zones
In order to focus more clearly on the kinds of trading zones that service scientists,
managers, and engineers will facilitate, let us consider a taxonomy of trading zones,
which can be classified into four broad states based on the extent to which they are
homogeneous or heterogeneous, and on whether they involve coercion or collabora-
tion (see table 5.1).18 The two classifying dimensions are on a continuum, and trading
zones can shift over time, becoming more or less heterogeneous and more or less
collaborative.
In table 5.1, the class of relatively equal trading zones is represented by the upper
right cell, where heterogeneous participants from different fields trade knowledge by
using a creole, or through an interactional expert who serves as a trade agent, or by
means of a boundary object, leading to three subtypes:
• A trading zone coordinated primarily by a creole: In this subtype, a creole gradually
emerges without the need for an interactional expert to facilitate trades.
• A trading zone coordinated primarily by interactional expertise: In this subtype, one

or more interactional experts facilitate trades among expertise cultures that may even-
tually develop a creole. Consider, as an example, the role played by AIDS treatment
activists in the testing of AZT (Epstein 1996). The activists gradually acquired enough
of the language of medical science to contribute to the design of new clinical trials.
But there was no creole—the scientists and activists did not evolve a new language.
In order to broker trades, interactional experts will have to understand not only the
content of another domain but also its perspective—the way it frames the problem.
The interactional expert will also have to be trusted by all parties. Unfortunately, the
acquisition of interactional expertise in science cost the AIDS activists credibility with
other parts of the AIDS community.
• A trading zone coordinated primarily by means of a boundary object, like a cowrie
shell or the Mars Polar Lander. The development of standard reference materials is a
good example. Organizations like the National Institute of Standards have to consult
the community in order to create a standard form of a material to be used as a refer-
ence for those engaged in research and production—so that different studies on, say,
the toxic effects of carbon nanotubes can actually be compared. Here the nanotubes
are a boundary object that can be represented differently by different groups, depend-
ing on how they prepare the materials and what they use to visualize their
structure.19
Obviously, there are cases where a trading zone will involve combinations of the
above: interactional expertise can be combined with the development of a creole and
with increasing alignment around the characteristics of a boundary object.
The top left cell of table 5.1 shows a situation where a trading zone has gradually
morphed into a new expertise with its own language and with those who were in
different expertise communities now seen as contributing to this new one. Here com-
Table 5.1
A Framework for Classifying Trading Zones
Homogeneous Heterogeneous
Collaboration Community: Equal trading zone:

interlanguage creole
Contributory expertise Interactional expertise
Shared paradigm Boundary object
Coercion Dominant ideology Enforced
Users converted to Users forced to conform to
designer’s mental model designer’s mental model
munity members will have varying degrees of inclusion in a shared paradigm, or form
of life, or mental model. The end result can be a new discipline, such as SSME might
become.20 Or it can be a new relationship between a producer and a client, leading to
coevolution of a new technology.
The likely pattern of evolution in table 5.1 is from the right cell to the left, although
it is possible for an existing disciplinary community to fragment into subspecialties
that form their own cultures, thus going from left to right. SSME is currently trying
to make this move from right to left, from a collection of trading zones struggling
toward a creole, to a recognized disciplinary community with degrees and programs
that produce experts.
The bottom two cells show situations where an elite participant dominates the zone
and there really is no trade. Centralized agricultural schemes favored in the Soviet
Union and other communist states are an example; these schemes were poor at agri-
culture, but good at controlling the population (Scott 1998). The state used both ideo-
logical and military means to enforce its top-down schemes.
On the bottom right, an elite enforces its dominance by force, economic power,
and/or military coercion. Poverty, as Amartya Sen (1999) has noted, is a kind of
“unfreedom.” Slavery is an extreme example; the slaves are free to have their own
ideology, as long as they do not rise up against their masters. Technology can
also constitute a kind of unfreedom, for example, by forcing users to conform to a
designer’s mental model, or in more extreme cases, structuring social relationships
in ways that particularly disadvantage certain groups.21
On the bottom left, the ideological element is singled out. Ideology is usually com-
bined with coercion, but does not have to be—once an ideology is internalized, its
adherents can carry on in the absence of threats. An example from a technological
system would be users who are converted to a designer’s mental model, perhaps after
years of use; think of the gap between Windows and Mac users, who often appear to
inhabit incommensurable worlds.
Less extreme cases illustrate lesser degrees of coercion or enforcement that allow
more trading. Consider how programs designed to help the poor worldwide can
involve coercion, if the aid is directed toward metrics designed to satisfy elite donors
and governments, rather than the needs of those who are suffering.22 Here the expert
is seen as having the solution, and the poor need to adopt it, as opposed to more
bottom-up approaches like that adopted by the Grameen Bank in Bangladesh, which
provides the locals with credit so they can design their own solution to poverty.
The lesson for SSME is that it may be useful to view a designer-user relationship,
or a producer-client relationship, as a trading zone, in which users and clients have
different types of expertise and operate from different paradigms. For example, service
scientists, managers, and engineers could experiment with more approaches that
matched the needs of the poor in specific locations with those who could offer assis-
tance, including assistance that came in the form of new technologies. But these
technologies would work best if the potential users were involved in the design. The
result would be a true trading zone between those with needs and those who could
supply the things needed.
These relatively equal trading zones frequently include a boundary object or system
that participants are trying to work together to create, and that sits on the boundaries
of their various types of expertise. In the case of radar, for example, different partici-
pants in the zone had partly unique perceptions of the emerging technology and its
potential (Galison 1997). Each expertise culture participating in the trading zone views
the boundary system from the standpoint of its own paradigms.
None of the existing literature on trading zones has probed the mental models of
different participants in a boundary system trading zone. Here service science could
make a genuine contribution to the literature, one akin to the breakthroughs made
on understanding designers’ and users’ mental models (Gorman 1995; Norman 1983;
Rouse and Morris 1986). For example, Donald Norman discusses the way in which
the designer’s and users’ mental models of common devices like computers and VCRs
can be far apart (Norman 1993). A VCR was thus a boundary object viewed very dif-
ferently by designers and users. To the designer, it was obvious how to program it; to
the user, it was impossibly difficult and arcane. Designers may have a particular script
they intend users to follow (Schot and Bruheze 2003), and if the users are not given
an opportunity to alter the technological system, or at least subvert it, the result is
coercion by the dominant ideology, the zone in the bottom left cell of table 5.1.
From an SSME perspective, designers and users need to coevolve technological
solutions.23 Their mental models of the boundary system need not be identical; the
designer will see technical aspects of the device that most users can ignore in the short
term, and the user will do things with the device that the designer could not antici-
pate. That is why these communities have to continue a dialog over the boundary
device. As the system of which the device is a part evolves, the trading zone will have
to morph as well—into new technologies like DVDs, which open new possibilities.
This example is an oversimplification, of course. There will be many different tech-
nical and user communities involved and other partners, such as the movie and music
industries. All of these negotiations will affect the kinds of artifacts that are produced,
as Bijker has shown in the case of the bicycle (Bijker 1995). The SSME will be working
as part of such complex sociotechnical networks.
On the cutting edge of coevolution, the service scientist and the client should
end up sharing a mental model of the client’s need and working together so closely
that they become collaborators. In this case, there is no incommensurability at the
level of system goals—even if the goal cannot be clearly stated because it is con-
stantly evolving, based on an increased understanding of the possibilities for the
system.
This state is most characteristic of cutting-edge design teams that created systems
like the ARPANET (Hughes 1998) and the Sidewinder missile (Westrum and Wilcox
1989). This state can also exist in trading zones that are not aimed at a joint system
design but have a shared understanding goal of another type. In all cases, there is
enough of a shared mental model of the system and/or goal that participants can work
toward it on their own without constant supervision, and can help each other evolve
a better system design and/or set of goals. Note that there can still be a division of
labor in a shared mental model zone, but that the organization is not rigid or hierar-
chical; roles evolve with the project, and participants move across task boundaries,
“flowing to the work” (Fisher and Fisher 1998). Parts of the boundary system are still
represented differently by those with unique specializations, but these types of exper-
tise are aligned around a common goal by continuous discussion.
An example is the team of physicists at Los Alamos that was at the center of atomic
bomb design and testing. From the start, the military wanted the project compart-
mentalized, with individual scientists knowing little or nothing about what was going
on in other parts of the division of labor. The scientists resisted, on the grounds that
knowledge sharing across the emerging organization was essential. A colloquium series
served as a kind of marketplace where ideas from any quarter could be aired and
exchanged. The usual hierarchical distinctions among scientists at different levels of
eminence disappeared, and disciplines had to be combined to address new problems
like how to build an implosion device. The director at Los Alamos, J. Robert
Oppenheimer, is described as
integrat[ing] the laboratory by his physical circulation through it, visiting meetings in theo-
retical physics, experimental physics and metallurgy. … Some commentary, indeed, ascribes
Oppenheimer’s skill at integration to the circumstance that he just knew an enormous amount
of the relevant physics and, more generally, that he had a grasp of a greater range of sciences
than anyone else at Los Alamos. He could walk into a technical discussion in an area about
which he might be presumed ignorant and make a decisive intervention, if not because of
his factual or theoretical knowledge, then because of his ability to cut incisively to the heart
of any problem. (Thorpe and Shapin 2000, 573)
Oppenheimer thus served as an interactional expert.24

Oppenheimer’s efforts to facilitate were helped by the fact that his community of
experts was relatively isolated at Los Alamos, and united by a common sense of
purpose—by the feeling that their work could save their country, even what they
perceived as their civilization. Service science should indeed be applied to the chal-
lenges faced by civilization. SSME experts may eventually be recruited into the security
area to deal with military and terrorist threats. Service scientists should also take the
lead in creating trading zones to deal with other global problems and opportunities,
like poverty and the environment. For example, Mehalik and Gorman (2006) have
described how an interdisciplinary design team worked together to invent a new kind
of environmentally intelligent textile. The atom bomb and textile teams, like the
ARPANET case, constitute proof of concept that it is possible for trading zones to shift
into homogeneous and collaborative communities in order to reach a common goal.
In these cases, the new community may be temporary; it might not spawn a new
discipline. Participants may return wistfully to their previous communities: “when the
Manhattan Project at Los Alamos was wound down, many who then returned with
relief to their normal university employments recalled that they had never had so
much fun and that science had never been freer” (Thorpe and Shapin 2000, 546).
These shared mental-model experiences transition between the top right and left
cells of table 5.1. The result can be a new expertise such as the community devoted
to the testing and maintenance of nuclear weapons that grew out of the Los Alamos
experience.25
For SSME, the lesson is that coevolution will require designers and users, producers
and clients to form a shared mental model of what they are trying to achieve, and to
complement that with interactional expertise sufficient to share strategies. Out of
these sorts of collaborations a new SSME expertise should emerge.
Creating IT That Facilitates Trading Zones: A Reflexive Application of SSME
SSME practitioners could work on coevolving new technologies to facilitate designer-

user collaborations. In 1960, J. C. Licklider, one of the architects of the revolutionary
changes that turned computing into information technology, proposed symbiosis
between human beings and computers as the eventual goal, in which the computer
becomes “a colleague whose competence supplements your own” (Waldrop 2001,
176). Symbiosis implies coevolution between computational technologies and human
cognitive and biological processes. Service scientists are ideally placed to facilitate
this coevolutionary process—and to monitor it, evaluating the results of multiple
experiments.
Being able to supply information when a user or team of users needs it depends on
having enough of a shared mental model of a problem to anticipate when the next
step differs from context to context in a way that cannot be entirely predicted by a
set of rules. Especially in a team situation, computational technologies will have to
facilitate sharing transactive memory, or the knowledge of who knows what (Wegner
1995). This knowledge evolves as a team works together, and a truly autonomous set
of computing aids would have to evolve with the group, becoming so embedded that
it may at some point be impossible to tell whether specific aspects of knowledge reside
primarily in brains, interactions, or information technology.
Another route is to evolve computational tools that serve as both boundary objects
and interlanguages in trading zones. Cognitive scientist and software engineer Jeff
Shrager (see his chapter in this volume) gained contributory expertise in molecular
biology (Shrager 2005)—and found that his colleagues in the biology laboratory
wanted to take advantage of his IT skills (Shrager 2007). He ended up doing more
programming than laboratory work. In order to get back to biology, Jeff started devel-
oping computational tools that the biologists themselves could program, and created
a trading zone where the computational tool served as a kind of interlanguage. He
began by creating a BioLisp, then one of his colleagues created BioLite. Each of these
tools contained new nouns and verbs that combined biological terms with software
functionality. Shrager explains: “Just as Galison considered physical instruments as
components of inter-languages, and as computers are now fundamental scientific
instruments, I consider programs, programming languages, and programming plat-
forms as representing components of the inter-languages that facilitate trade between
collaborating biologists, and especially between [biologists] and software engineers
who program the computers to serve biology in this capacity” (Shrager 2007, 645).
The current iteration of these tools is called BioBike:
The programmable client-server architecture of BioBike enables both the programmers and biolo-
gists to evolve the programming language. This is an important difference—not only are pro-
grammers extending the language, but biologist users are as well, usually with some assistance
from programmers. In order for this to be effectively organized, the biologists have to learn
something about programming, and the programmers have to learn something about biology—
and, of course, they need to learn enough to collaborate with one another. In the BioBike com-
munity the various participants must at least learn to the level of having “interactional expertise”
with regard to one another. (Shrager 2007, 646)
In other words, biologists and programmers are coevolving new solutions to emerging
problems and opportunities. BioBike serves both as an interlanguage and a boundary
object, connecting a trading zone that is a model for SSME.26
Information technology can also be used to facilitate trading zones across a wider
range of stakeholder groups. Arizona State University’s Decision Theater allows mul-
tiple stakeholders to visualize the impact of their assumptions on future system states.
For example, different growth scenarios for Phoenix can be projected in three dimen-
sions to facilitate a kind of visual creole that will allow stakeholders to see what
happens when their values are translated into policy.
Advances in real-time visual communication could create virtual decision theaters
linking stakeholders across the globe—especially those in the remotest and poorest
areas, who never get to sit at the conference table with the rich and powerful. This
kind of technology is a perfect application area for SSME because it will evolve based
on the interactions between designers and users, who could prototype new designs
and management strategies for complex systems in a virtual environment.
When Participants Disagree on the Type of Trading Zone
Intellectual property issues can threaten the co-creation relationship sought by service
scientists, and prevent STS practitioners from studying this process. There are ways of
navigating these difficulties with nondisclosure agreements, but even these depend
on a relationship of trust.
Consider the partnership between Lexar and Toshiba to develop new forms of
controllers for flash memory in cameras (Gorman 2006). Lexar is a small company
that relies on design innovations, while Toshiba is one of the largest manufacturers
of NandFlash memory. Lexar could supply controller designs that would increase the
market for the NandFlash memory produced by Toshiba.
The Toshiba-Lexar relationship foundered on a misperception of what kind of
trading zone was involved. Lexar saw the relationship between the two companies as
a strategic partnership, based on trust, and shared all of its information with its
partner. In contrast, Toshiba saw the relationship as a trading zone, bounded by legal
agreements that required the parties to specify exactly what knowledge was restricted.
In terms of the framework we are developing here, Lexar believed it and Toshiba were
in a shared mental model state (in the upper left cell of table 5.1), and Toshiba saw
the relationship as a more standard trading zone (in the upper right of table 5.1). This
mismatch in perceptions was one of the sources of difficulties between the two com-
panies, and brought their nascent trading zone to a halt. Toshiba engineers who
worked with Lexar also worked with Lexar’s competitors who were developing control-
lers. When Lexar found out, they sued, and eventually won a settlement. Service
scientists and managers will have to be adept not only at managing the legal aspects
of intellectual property, but also the type of trading zones partners want to establish—
making sure that perceptions as well as agreements are aligned.
Ultimately, a shared mental model can lead to a collaboration that is no longer a
trading zone. In the case of a service, the result would be a blurring of the provider
and client roles, as both parties become so engaged in the coproduction of value that
they no longer worry about who gets credit for what. Unfortunately, the usual result
of such a situation is that the parties involved look back afterward and think each got
too little credit, and then recriminations ensue—and sometimes even litigation, as the
example above indicates. Trust is a critical component in effective trading zones, and
no amount of legal agreements will substitute for it (Seabright 2004).
Ethical Issues in SSME
Imagining the future of information technology requires explicit consideration of

values—that is, what kind of world we ought to create. The sacred books and mytholo-
gies that date back hundreds of years cannot account for the dilemmas and capabilities
created by advances in science and technology.27 Autonomous and pervasive informa-
tion technologies (Waldrop 2003), coupled with biotechnology and nanoscience, will
take us closer to abilities reserved for God(s) in most of our traditional stories. There-
fore, decisions about which technological directions deserve public support, both
nationally and internationally, will involve the exercise of moral imagination (Werhane
1999).
A service scientist may be confronted with situations in which a client wants to do
something unethical—for example, using IT in a way that invades privacy without
consent. Service scientists, like other professionals, will need an ethical code that
reminds them of their moral obligations. Codes are useful guidelines, but need to
evolve with the technological frontier and with changes in the roles of service scien-
tists. Codes need to be complemented by moral imagination—and moral imagination
will require the development of trading zones that involve not only client and pro-
vider, but also stakeholders likely to be affected by the new technology. STS scholars
could play a role in SSME trading zones, helping to identify and recruit key
stakeholders.
Organizations and Trading Zones
So far, our examples of trading zones have consisted mainly of those teams assembled
for a task or purpose—creating radar, conducting a workshop, or applying for a major
grant. Obviously, service scientists also will form trading zones that exist within and
across organizations. These trading zones can be ways of working around the restric-
tions imposed by the formal organizational structure.
Organizations can fall into patterns of normalized deviance. For example, NASA’s
decisions about flying each of the space shuttles involved a complex bureaucratic
process (Vaughan 1999). The hierarchical and compartmentalized organization of
NASA prevented trading zones that would have linked expertise communities across
multiple levels of the organization. What lower-level engineers regarded as a signifi-
cant risk was normalized into an acceptable risk by the bureaucratic process. According
to Richard Feynman, who served on the Presidential Commission evaluating the Chal-
lenger disaster, “It appears that there are enormous differences of opinion as to the
probability of failure with loss of vehicle and human life. The estimates range from 1
in 100 to 1 in 100,000. The higher figure comes from working engineers, and the very
low figures from management” (quoted in Vaughan 1996, 274). Feynman himself felt
the probability was about 1 in 25; as the later Columbia disaster illustrated, his estimate
was unfortunately closer to the mark.
The debacle at Enron is another example of normalized deviance. This former
pipeline company became a trading company, which might have been fine, but cre-
ative accounting was used to make it seem as though the trades were recording a profit,
when in fact most were losing money. Enron used mark-to-market accounting for
every deal, which means it booked the expected long-term revenue for each deal
immediately. Note the term “expected.” Enron traders could make up a figure for the
long-term revenue, and have it count in the current quarter. Traders got paid cash
bonuses based on this estimate of the value of the deal, so that their emphasis was
on making deals, not on making money.
When the company was imploding under its debt and investigations were under
way, a new treasurer was appointed. He asked about cash flow, and learned that there
was no method for tracking it. He asked about the debt schedule, and found Enron
was not tracking when its debts were due. As Kurt Eichenwald reports: “With all the
focus on deals and earnings—with the finance group’s transformation into a profit
center rather than a division to support the business—the workday, boring details had
been sloughed off. … Enron had been flying blind financially for years” (Eichenwald
2005, 560–561). It took a new treasurer to see that what passed for normal accounting
at Enron was totally deviant—it broke the basic rules of Finance 101.28
Normalized deviance occurs when practitioners within a paradigm create a set of
procedures and roles that allow them to dismiss or cover up anomalies. Trading zones
by definition involve different paradigmatic communities. Therefore, when properly
constructed, trading zones could help prevent normalized deviance by bringing an

outsider’s perspective to a client’s problem. The service scientist should not accept the
client’s definition of the problem, but should follow one of the maxims of systems
engineering and try to determine what the real problem is (Gibson, Scherer, and
Gibson 2007). What Enron’s management wanted was to push up its stock price; what
they needed was a system to continuously assess which parts of the company were
actually making money.
Service scientists can facilitate the development of trading zones within organiza-
tions in order to figure out what a client really needs, as opposed to what they want.
Simply working with one part of the organization is not enough, especially in hierar-
chical organizations where decision making is compartmentalized. The service scien-
tist may have to act like an anthropologist, ferreting out relevant expertise and
bringing it together to coevolve new solutions.
Is SSME an Esoteric or an Interactional Expertise?
Do any of the SSME opportunities discussed above require a new form of contributory
expertise, or will interactional expertise suffice? Contributory experts in a domain are
able to:
1. Classify certain problems in ways that suggest their solution, so that recognizing a
problem is almost the same as solving it.29 This is akin to Kuhn’s normal science; the
scientist working within a paradigm knows what problems are worth solving, and also
what tools can be used to solve them.
2. Develop a toolbox of heuristics and technological capabilities that will help them
solve problems, once they are classified.
3. Share a common language that allows the experts to quickly exchange information,
using a jargon that is hard for those outside the community to penetrate without
gaining interactional expertise.
4. Develop professional organizations, conferences, degree programs, and other evi-
dence of a professional identity.
Service science is already doing the fourth step, which is the most important one, if
the point of an organization is to facilitate training and the sharing of knowledge
among those who practice a kind of SSME art, and to convince others that this kind
of skill is useful. This kind of expertise has an important experiential component,
similar to what a good therapist or business consultant does—where experience
obtained from multiple case situations is as important as the formal training.30 But
this kind of art is unlikely to evolve into a science unless it is reflexive—unless SSME
studies and tries to codify its own knowledge.
Another problem with labeling service a science is that the SSME will not be offer-
ing packaged solutions; instead, the SSME expert must always be on the alert for
creative reformulations of a problem, in collaboration with a client. In other words,
SSME ought never to be in normal science mode.
Conclusions and Future Directions
The central question raised by this article is whether SSME will evolve into a distinct
expertise community, or whether it will remain primarily an interactional expertise.
To find out, researchers interested in expertise and experience can focus on the
following questions:
• Whether trading zones emerge, and if so, and what kinds they are: Are they medi-
ated by interlanguages? by interactional expertise? by some combination? Are there
interdisciplinary interactions that cannot be described as trading zones?
• What kinds of shared terms and meanings evolve? For example, how is the core idea
of a combination of service science and management represented by different con-
stituencies at different levels of inclusion in the emerging technological frame?
• What people consider themselves experts in SSME? How do they operate?
• What sorts of educational and certification programs emerge? Do textbooks begin
to develop a “normal science” account of the field?
STS can help extend the boundaries of services to include those who are usually not
considered “clients” in the business sense. SSME requires what systems engineers refer
to as outscoping, or determining what a client really needs—which may be different
from what they say they want (Gibson, Scherer, and Gibson 2007). At least some
service scientists should be outscoping on a global level, thinking about issues such
as raising the standard of affluence and improving the environment.31
In his work on poverty and development, Easterly (2006) has called for locating
“seekers” around the world who want the chance to coproduce value, but are limited
by lack of initial resources and infrastructure. SSME can help raise the global level
of affluence and the level of education and freedom for women, both key to con-
trolling population growth (Sen 1999) and making armed revolution less attractive.32
A good example is Greg Mortenson’s work building schools in Pakistan: in order
to provide an essential service, especially to women, he had to create a trading
zone that linked seekers in remote areas with the religious establishment, with the
national government, and with donors in the developed world (Mortenson and
Relin 2006).
SSME can also play a critical role in ensuring that new technologies represent social
as well as technical progress. Indeed, if these technologies are to produce value and
revenue, they have to represent social progress. Building new systems at the conver-
gence of nano-, bio-, information, and cognitive technologies will require very close
and repeated interactions among multiple seeker groups (Gorman and Groves 2006).
SSME demands cooperation, collaboration, and humility.33 Every service scientist will
end up being a reflective practitioner (Schon 1987), seeing not just the system but also
her part in it. Cognitive diaries are a good tool for this kind of reflection (Shrager
2005).
Service scientists will need training in a core discipline—computer science, cogni-
tive science, environmental science, medicine, or law, depending on the type of
systems they intend to specialize in—though the boundaries between systems are
fuzzy at best, and do not correspond to traditional disciplines. The interactional com-
ponent will require every service scientist to gain skills in facilitating and managing
trading zones, a new kind of competence that will draw on disciplines like anthropol-
ogy and social psychology, but will move beyond what is currently known. Inevitably,
such training will have an experiential component, in which service scientists serve
as apprentices to those more experienced. If the stakeholder net is drawn widely
enough, SSME could potentially benefit large portions of the global population that
are not usually seen as constituting “business opportunities.”
Whether or not an esoteric expertise evolves, SSME will be a domain that requires
interactional expertise, and is a worthy object of study for STS. The SEE framework
discussed in this paper is one way of approaching such a study, which should lead to
improvements not only in service science but also in the framework itself.
Notes
1. See “Succeeding through Service Innovation: Developing a Service Perspective on Economic

Growth and Prosperity,” Cambridge Service Science, Management, and Engineering Symposium,
Cambridge University, July 14–15, 2007 (www.ifm.eng.cam.ac.uk/ssme), 7.
2. For the latest information, see IBM’s SSME site, http://www.almaden.ibm.com/asr/SSME/. For
an example of a university working on SSME, see http://ssme.berkeley.edu. For an example of a
company working on SSME, see http://www.hpl.hp.com/news/2005/oct-dec/services.html, and
on the Service Research Innovation Initiative, see http://www.thesrii.org. For an example of a
nation working on SSME, see http://www.ssme.in. Also, a 2007 U.S. Innovation Initiative act (HR
2272, Section 1106) calls for “The Study of Service Science.”
3. See the NSF Web page on “Service Enterprise Systems” at http://www.nsf.gov/funding/

pgm_summ.jsp?pims_id=13343&org=NSF.
4. Doug Morse of Oracle discussed this problem at the sixteenth annual Frontiers in Service
conference in San Francisco on October 6, 2007.
5. See Hira and Goldstein 2006: “The [U.S.] National Competitiveness Investment Act, intro-
duced in September by a whopping 38 senators from both political parties, includes a
section supporting government initiatives in services science. … What is services science?
The legislation defines it vaguely as ‘curricula, training, and research programs that are
designed to teach individuals to apply scientific, engineering, and management disciplines
that integrate elements of computer science, operations research, industrial engineering, busi-
ness strategy, management sciences, and social and legal sciences, in order to encourage
innovation in how organizations create value for customers and shareholders that could not
be achieved through such disciplines working in isolation.’ That’s not a definition, just a
laundry list.”
6. For a periodic table of expertise, and more details on stages in the acquisition of various catego-
ries of expertise, see chapter 4 of this volume and Collins and Evans 2007.
7. A good example of interactional expertise is the American special forces soldier in Yemen who
mastered the culture of Qat to the point where he could carry out sophisticated arguments about
which kinds of Qat were preferable without ever having chewed it. See Kaplan 2005.
8. Based on a presentation by Dave Kreutz at the American Association for the Advancement of
Science, 2007 (Kreutz 2007), and conversations with him afterward.
9. See Gorman and Mehalik 2002 for examples of service situations such as getting power to
villagers in rural South Africa and Vietnam.
10. One anonymous reviewer of a draft of this paper commented that the interactional expert
sounds a lot like a heterogeneous engineer. The latter designs and/or manages systems that
combine heterogeneous elements, such as the combination of natural, technological, and
human factors involved in Portuguese navigation (Law 1987). In fact, the interactional expert
possesses a kind of expertise that would be useful in doing heterogeneous engineering, because
typically more than one expertise is involved. Dave Kreutz, for example, has to navigate a
heterogeneous combination of elements; his interactional expertise helps him with the bycatch
reduction part of the system he is managing.
11. IBM at one point characterized technical knowledge as the vertical bar and business knowl-
edge as the horizontal, indicating that the T-shaped concept is more amorphous than interac-
tional expertise (Glushko 2008).
12. “Succeeding through Service Innovation,” 12.
13. This lack of training in interactional expertise is a problem for the growing convergence
across nano-, bio-, information, and cognitive technologies; see Gorman 2004.
14. “Two men who perceive the same situation differently but nevertheless employ the same
vocabulary in its discussion must be using words differently. They speak, that is, from what I
have called incommensurable differences” (Kuhn 1970, 200).
15. Kuhn recognized the importance of working toward at least a common jargon when
crossing paradigm boundaries: “what the participants in a communication breakdown can do
is recognize each other as members of different language communities and then become
translators. Taking the differences between their own intra- and inter-group discourse as itself
a subject for study, they can first attempt to discover the terms and locutions that, used
unproblematically within each community, are nevertheless foci of trouble for inter-group
discussions” (Kuhn 1970, 202). As an example, Galison (1997) describes Tim Dutton’s 1985
analysis of the development of “Police Motu” in what is now Papua New Guinea. The Motu
developed a simplified version of their language to facilitate trade across their extensive
network. This reduced version of Motu became a pidgin when the British adapted it to colo-
nial rule, introducing new terms and uses that stabilized. When the pidgin is elaborated
sufficiently for children to grow up in it, so to speak, it becomes a creole—like the dominant
language in Haiti, which began as a French-based pidgin used for communication both between
colonizers and subjects and also among the different linguistic groups enslaved by the French
(see Asher 1994). In his discussion of the collaboration to develop radar and particle accelera-
tors, Galison notes, “the most effective feature of the huge war laboratories was the imposed
orchestration of the practices of theorists, experimenters and instrument makers with those
of electronic and mechanical engineers” (Galison 1997, 837). This “imposed orchestration”
resembles the need for the British colonizers to communicate with the Motus, and for slaves
from a variety of cultures to understand the French and each other. Pidgins are especially
likely to develop and survive to the creole stage when three or more languages intersect
(Todd 1990).
16. Participants in this situation actually referred to their negotiations as a “trade” (Lambert and
Shaw 2002). For the JPL participants, a trade was very different from a tradeoff; the former referred
to choosing one out of a set of parallel design alternatives, whereas “tradeoff” referred to balanc-
ing different technical parameters in an overall design.
17. “Services exchange … involves a negotiated exchange between a provider and an adopter
(supplier and customer) for the provision of (predominately) intangible assets. … When products
were the main focus, the information they contained helped each side communicate effectively
with the other. As products and their functions became better understood, suppliers did not need
to understand the customer’s business to be an exchange partner. Similarly, customers did not
need to understand their suppliers’ prior experiences and capabilities, since these were reflected
in the products they could see, touch and experience directly. The services transaction is differ-
ent. The exchange is generated by both parties, and the process of adoption or consumption is
an integral part of the transaction. So, the adopter or customer is also a co-producer, intimately
involved in defining, shaping and integrating the service into his or her organization”
(Chesbrough 2004, 1).
18. This four-state, two-axis taxonomy of trading zones was developed by Collins, Evans, and
Gorman (2007); it supplants a previous three-state framework developed by Gorman and Mehalik
(2002). Application of this framework to service science will create opportunities for empirical
research that will lead to its modification and improvement.
19. A standard reference material can serve to align the groups into a productive trading zone.
This example comes from the senior author’s participation in a workshop on Standards for
Environmental, Health and Safety Needs for Engineered Nanoscale Materials (National Institute
of Standards and Technology, September 12–13, 2007), where the meanings of reference materials
and standards for different communities were discussed.
20. The creation of disciplines depends both on an expertise that can be differentiated from
that of other fields and establishing enough social legitimacy so that those trained in the new
field will be able to find meaningful employment (Hambrick and Chen 2008). One way to
establish legitimacy is to show how the paradigm of a new field emulates that of one or more
well-established areas. For SSME, this creates a tension: the new field could gain greater legiti-
macy by being linked to existing fields like marketing and computer science, but it needs to
establish a unique expertise that cannot be obtained from any of the many fields that feed
into SSME.
21. See, for example, Winner 1986 and the discussion of “subversive trading zones” in chapter
2 of this volume.
22. Easterly (2006) makes the argument that poverty aid is usually directed by top-down planners
working in an extensive bureaucracy, who assume that they know best—that they can redesign
other countries and situations. What is needed is to take advantage of local aid seekers who want
a service or product, and match them with both providers and funding.
23. Recent STS work has highlighted the way in which users and designers co-construct technolo-
gies. The user is not a passive recipient of technology; see Oudshoorn and Pinch 2003. Coevolu-
tion emphasizes the way in which co-construction can be a process that takes place over time,
and in the case of SSME, is deliberate: the expectation is that technology should evolve out of
repeated negotiations between designers and users, and that at some point these roles may blur
as they work together to coproduce value.
24. Thorpe and Shapin (2000) claim that Oppenheimer achieved a kind of interactional
charisma, showing that his influence was not due solely to his expertise. I. I. Rabi, working at
MIT’s radiation laboratory, tutored Oppenheimer in how to deal with the military; thus, the
Los Alamos environment encouraged the development of knowledge trading zones that would
have been prevented by the military’s obsession with compartmentalization.
25. Mackenzie and Spinardi (1995) have discussed how this expertise community gained signifi-
cant tacit as well as explicit knowledge about the testing and maintenance of nuclear weapons.
26. Similarly, David Wallace and colleagues at MIT have created a Web-based computational
tool that allows for intertranslation of different models of a system—a kind of interlanguage that
facilitates coordination. For example, the University of Tokyo was involved in creating a virtual
model of Tokyo that involves multiple models of different systems and areas—models that would
ordinarily be difficult to combine. Wallace’s software makes this combination process relatively
easy for a user (Borland and Wallace 2000).
27. Joseph Campbell makes this point in multiple lectures and writings (see, e.g., Campbell
1968). Donna Haraway’s cyborg is a step toward a new kind of story, or identity, that might fit
our current human situation better than the old stories (Haraway 1991).
28. Here Seabright’s metaphor is apt: “Just as the sensory capacities of a large mammal enable
it to see, or hear, or scent danger before the danger arrives, and thereby check the urge to act
on the impulses of hunger or libido, so the checks and balances of a modern corporation are
means to sense disaster before it arrives, and thereby restrain the urge to act on an impulse, even
a great and creative impulse. When a corporation’s internal checks begin to fail, its days are as
surely numbered as those of a large hunting mammal whose sense or hearing fades. And the
larger it is, the faster it will run out of fuel for its daily metabolic needs” (Seabright 2004, 172).
29. For an overview of research on expertise, see Ericsson 2006. Classic articles relevant to science
and technology include Chi, Feltovich, and Glaser 1981; Larkin 1983; Gorman 1997; and Alberdi,
Sleeman, and Korpi 2000.
30. See Klein 1999; Klein 1989. This kind of experiential expertise may eventually be able to be
partly codified, typically by a leading figure, or guru, who establishes a school of therapy—like
psychoanalysis—or a consulting business based on her or his principles. If SSME went in this
direction, it would be fragmented into different styles or types associated with high
practitioners.
31. In the environmental area, Brad Allenby is independently advocating a new kind of expertise
that has many of the same characteristics as SSME. Allenby calls his new field earth systems
engineering management (ESEM); see Allenby 2005. Every part of the globe is now affected by
human activity; therefore, our species has a responsibility to monitor and manage our interac-
tions. Smaller, apparently more local environmental problems can no longer be considered in
isolation. Management of the Everglades, for example, will be affected by global warming, which
could bury much of this delicately balanced system under saltwater. The earth systems engineer
may work as a consultant on the design of complex systems like the Everglades, using adaptive
management techniques (Light, Gunderson, and Holling 1995).
The growing service sector of the global economy poses a similar set of problems. Changes
in one part of this growing global network may have unanticipated ripple effects in others.
Allenby (2005) proposes developing an ESEM expertise to facilitate management of the global
ecosystem. Similarly, managing the service economy requires a new kind of expertise in service
science, not unlike the broad, system-level expertise that Allenby proposes.
ESEM and SSME are not formally linked and evolved independently. Both illustrate the need
for combining multiple disciplines to form new specialties that increase our ability to manage
the way we are transforming the sociotechnical systems of which we are a part. ESEM might
even evolve into a subspecialty of SSME. Like SSME, ESEM is worthy of careful study, using the
framework outlined in this paper and other STS approaches.
32. People who become deeply interdependent through exchanges of services might be less likely
to go to war, though trade has traditionally not prevented military conflict between neighbors.
“The problem of civilized society … is how to turn the propensity to truck, barter and exchange
into something stronger than a propensity—into a habit, into second nature” (Seabright 2004,
234).
33. One of the major stumbling blocks is the extent to which results of these coevolutionary
experiments can be widely shared. SSME interactions will presumably result in the development
of new intellectual property, which cannot be shared without a variety of legal and ethical pro-
tections (Gorman 2006). SSME work with seeker groups across the planet may require the devel-
opment of new approaches to intellectual property (Werhane and Gorman 2005).
References
Alberdi, E., D. H. Sleeman, and M. Korpi. 2000. Accommodating Surprise in Taxonomic Tasks:
The Role of Expertise. Cognitive Science 24 (1):53–91.
Allenby, Brad. 2005. Technology at the Global Scale: Integrative Cognitivism and Earth Systems
Engineering Management. In Scientific and Technological Thinking, ed. Michael E. Gorman, Ryan
D. Tweney, David C. Gooding, and Alexandra Kincannon, 303–344. Mahwah, NJ: Lawrence
Erlbaum Associates.
Asher, R. E. 1994. Haiti: Language Situation. In The Encyclopedia of Language and Linguistics, ed.
R. E. Asher, 1520–1521. Oxford: Pergamon Press.
Bijker, Wiebe E. 1995. Of Bicycles, Bakelites and Bulbs: Toward a Theory of Sociotechnical Change.
Borland, Nick, and David Wallace. 2000. A Collaborative Internet-Based Modeling Approach.
Journal of Industrial Ecology 3 (2–3):33–46.
Campbell, Joseph. 1968. The Hero with a Thousand Faces. Princeton: Princeton University
Press.
Chesbrough, H. 2004. A Failing Grade for the Innovation Academy. Financial Times, September
24, 1.
Chesbrough, H., and J. Spohrer. 2006. A Research Manifesto for Services Science. Communications
of the ACM 49 (7):35–40.
Chi, M. T. H., P. J. Feltovich, and R. Glaser. 1981. Categorization and Representation of Physics
Problems by Experts and Novices. Cognitive Science 5:121–152.
Collins, Harry. 2004a. Gravity’s Shadow: The Search for Gravitational Waves. Chicago: University
of Chicago Press.
Collins, Harry. 2004b. How Do You Know if You’ve Alternated? Social Studies of Science 34
(1):103–106.
Collins, Harry, and Robert Evans. 2002. The Third Wave of Science Studies. Social Studies of Science
32 (2):235–296.
Collins, Harry, and Robert Evans. 2007. Rethinking Expertise. Chicago: University of Chicago
Press.
Collins, Harry, Robert Evans, and Mike Gorman. 2007. Trading Zones and Interactional Expertise.
Studies in History and Philosophy of Science 39 (1):657–666.
Collins, Harry, Robert Evans, Rodrigo Ribeiro, and Martin Hall. 2006. Experiments with Interac-
tional Expertise. Studies in History and Philosophy of Science 37 (4):656–674.
Davis, M. M., and I. Berdrow. 2008. Service Science: Catalyst for Change in Business Curricula.
IBM Systems Journal 47 (1):5–14.
Easterly, William Russell. 2006. The White Man’s Burden: Why the West’s Efforts to Aid the Rest
Have Done So Much Ill and So Little Good. New York: Penguin Press.
Eichenwald, Kurt. 2005. Conspiracy of Fools: A True Story. New York: Broadway Books.
Elzen, B., and B. Enserink. 1996. Socio-technical Networks: How a Technology Studies Approach
May Help Us Solve Problems Related to Technical Change. Social Studies of Science 26
(1):96–141.
Epstein, Steven. 1996. Impure Science: AIDs, Activism, and the Politics of Knowledge. Berkeley:
University of California Press.
Ericsson, K. A. 2006. The Cambridge Handbook of Expertise and Expert Performance. Cambridge:
Cambridge University Press.
Fisher, Kimball, and Mareen Duncan Fisher. 1998. The Distributed Mind: Achieving High Perfor-
mance through the Collective Intelligence of Knowledge Work Teams. New York: Amacom.
Chicago Press.
Gibson, John E., William T. Scherer, and William F. Gibson. 2007. How to Do Systems Analysis.
Indianapolis: Wiley.
Glushko, R. J. 2008. Designing a Service Science Discipline with Discipline. IBM Systems Journal
47 (1):15–27.
Gorman, Michael E. 1995. Confirmation, Disconfirmation and Invention: The Case of Alexander
Graham Bell and the Telephone. Thinking and Reasoning 1 (1):31–53.
Gorman, M. E. 1997. Mind in the World: Cognition and Practice in the Invention of the Tele-
phone. Social Studies of Science 27 (4):583–624.
Gorman, M. E. 2004. Collaborating on Convergent Technologies: Education and Practice.

In The Coevolution of Human Potential and Converging Technologies, ed. M. C. Roco and C. D.
Montemagno, 25–37. New York: New York Academy of Sciences.
Gorman, Michael E. 2006. STS, Ethics, and Knowledge Transfer in the Courtroom. Social Studies
of Science 36 (7):861–866.
Gorman, Michael E., and James Groves. 2006. Collaborating on Converging Technologies: Educa-
tion and Practice. In Managing Nano-bio-info-cogno Innovations, ed. William Sims Bainbridge and
Mihail C. Roco, 71–88. Dordrecht: Springer.
Gorman, M. E., J. F. Groves, and R. K. Catalano. 2004. Collaborative Research into the Societal
Dimensions of Nanotechnology: A Model and Case Study. IEEE Technology and Society Magazine
23 (4):55–62.
Gorman, Michael E., Alexandra Kincannon, and Matthew M. Mehalik. 2001. Spherical Horses
and Shared Toothbrushes: Lessons Learned from a Workshop on Scientific and Technological
Thinking. In Discovery Science 2001, ed. Klaus Jantke and Ayumi Shinohara, 74–86. Washington,
DC: Kluwer Academic Press.
Gorman, Michael E., and Matthew M. Mehalik. 2002. Turning Good into Gold: A Comparative
Study of Two Environmental Invention Networks. Science, Technology and Human Values 27
(4):499–529.
Hackett, Edward. 2000. Interdisciplinary Research Initiatives at the U.S. National Science Founda-
tion. In Practising Interdisciplinarity, ed. Peter Weingart and Nico Stehr, 248–259. Toronto:
University of Toronto Press.
Hambrick, D. C., and M. Chen. 2008. New Academic Fields as Admittance-Seeking Social Move-
ments: The Case of Strategic Management. Academy of Management Review 30 (1):32–54.
Haraway, Donna. 1991. A Cyborg Manifesto: Science, Technology, and Socialist-Feminism in the
Late Twentieth Century. In Simians, Cyborgs and Women: The Reinvention of Nature, ed. Donna
Haraway, 149–181. New York: Routledge.
Hira, Ron, and Harry Goldstein. 2006. IBM Takes the Guesswork Out of Services Consulting: Big
Blue Pushes a New Research Discipline Called Services Science. IEEE Spectrum (December), http://
spectrum.ieee.org/computing/hardware/ibm-takes-the-guesswork-out-of-services-consulting/.
Hughes, Thomas P. 1998. Rescuing Prometheus. New York: Pantheon.
IfM and IBM. 2007. Succeeding through Service Innovation: A Discussion Paper. Cambridge, UK:
University of Cambridge Institute for Manufacturing.
Jenkins, Lekelia. 2007. Interactional Expertise and the Problem of Dolphins Killed in U.S. Tuna
Purse Seine Fishery. Studies in History and Philosophy of Science 39 (1):698–712.
Johnson, Mark. 1993. Moral Imagination. Chicago: University of Chicago Press.
Kaplan, R. D. 2005. Imperial Grunts: The American Military on the Ground. New York: Random
House.
Klein, G. A. 1989. Recognition-Primed Decisions. In Advances in Man-Machine Systems Research,

ed. W. B. Rouse, 5:47–92. Greenwich, CT: JAI Press.
Klein, Gary. 1999. Sources of Power: How People Make Decisions. Cambridge, MA: MIT Press.
Kreutz, David. 2007. Horizontal Communication: Changing Fishers Attitude and Culture. Paper
presented at American Association for the Advancement of Science, San Francisco.
Kuhn, Thomas S. 1970. The Structure of Scientific Revolutions. 2nd ed. Chicago: University of
Chicago Press.
Lambert, Monique H., and Ben Shaw. 2002. Transactive Memory and Exception Handling in
High-Performance Project Teams. CIFE Technical Report #137, http://www.stanford.edu/group/
CIFE/online.publications/TR137.pdf.
Larkin, J. 1983. The Role of Problem Representation in Physics. In Mental Models, ed. D. Gentner
and A. L. Stevens, 75–98. Hillsdale, NJ: Lawrence Erlbaum Associates.
Law, J. 1987. Technology and Heterogeneous Engineering. In The Social Construction of Techno-
logical Systems, ed. W. E. Bjiker, T. P. Hughes, and T. J. Pinch, 111–134. Cambridge, MA: MIT
Press.
Light, S. S., L. H. Gunderson, and C. S. Holling. 1995. The Everglades: Evolution of Management
in a Turbulent Ecosystem. In Barriers and Bridges to the Renewal of Ecosystems and Institutions, ed.
L. H. Gunderson, C. S. Holling, and S. S. Light, 163–168. New York: Columbia University Press.
Mackenzie, D., and G. Spinardi. 1995. Tacit Knowledge, Weapons Design, and the Uninvention
of Nuclear Weapons. American Journal of Sociology 101 (1):44–99.
Mehalik, Matthew M., and Michael E. Gorman. 2006. A Framework for Strategic Network Design
Assessment, Decision Making, and Moral Imagination. Science, Technology and Human Values 31
(3):289–308.
Mortenson, G., and D. O. Relin. 2006. Three Cups of Tea: One Man’s Mission to Fight Terrorism and
Build Nations—One School at a Time. New York: Viking.
National Academy of Engineering. 2003. The Impact of Academic Research on Industrial Performance.
Washington, DC: National Academies Press.
Norman, D. A. 1983. Some Observations on Mental Models. In Mental Models, ed. D. Gentner
and A. L. Stevens, 7–14. Hillsdale, NJ: Lawrence Erlbaum Associates.
Norman, Donald A. 1993. Things That Make Us Smart: Defending Human Attributes in the Age of
the Machine. New York: Addison Wesley.
North, Douglass Cecil. 2005. Understanding the Process of Economic Change. Princeton: Princeton
University Press.
Oudshoorn, Nelly, and Trevor Pinch, eds. 2003. How Users Matter: The Co-construction of Users and
Technologies. Cambridge, MA: MIT Press.
Palmer, Carole L. 2001. Work at the Boundaries of Science: Information and the Interdisciplinary
Research Process. Dordrecht: Kluwer Academic Publishers.
Paulson, Linda Dailey. 2006. Services Science: A New Field for Today’s Economy. IEEE Computer
(August): 18–21.
Ribeiro, R. 2007. The Language Barrier as an Aid to Communication. Social Studies of Science 37
(4):561–584.
Rouse, W. B., and N. M. Morris. 1986. On Looking into the Black Box: Prospects and Limits in
the Search for Mental Models. Psychological Bulletin 100 (3):349–363.
Schon, Donald A. 1987. Educating the Reflective Practitioner: Toward a New Design for Teaching and
Learning in the Professions. San Francisco: Jossey-Bass.
Schot, Johan, and Adri Albert de la Bruheze. 2003. The Mediated Design of Products,
Consumption, and Consumers in the Twentieth Century. In Oudshoorn and Pinch 2003,
229–245.
Scott, James C. 1998. Seeing Like a State: How Certain Schemes to Improve the Human Condition Have
Seabright, Paul. 2004. The Company of Strangers: A Natural History of Economic Life. Princeton:
Princeton University Press.
Sen, Amartya. 1994. Population: Delusion and Reality. New York Review of Books 41 (15):62–71.
Sen, Amartya. 1999. Development as Freedom. New York: Random House.
Shrager, Jeff. 2005. Diary of an Insane Cell Mechanic. In Scientific and Technological Thinking, ed.
Michael E. Gorman, Ryan D. Tweney, David C. Gooding, and Alexandra Kincannon, 119–136.
Mahwah, NJ: Lawrence Erlbaum Associates.
form. Studies in History and Philosophy of Science 39 (1):642–656.
Spohrer, Jim, Paul P. Maglio, John Bailey, and Daniel Gruhl. 2007. Towards a Science of Service
Systems. Computer 40 (1):71–77.
Spohrer, J. C., D. McDavid, P. P. Maglio, and J. W. Cortada. 2006. NBIC Convergence and
Technology-Business Coevolution: Towards a Services Science to Increase Productivity Capacity.
In Managing Nano-bio-info-cogno Innovations, ed. William Sims Bainbridge and Mihail C. Roco,
71–88. Dordrecht: Springer.
Thorpe, Charles, and Steven Shapin. 2000. Who Was J. Robert Oppenheimer? Social Studies of
Science 30 (4):545–590.
Todd, Loreto. 1990. Pidgins and Creoles. London: Routledge and Kegan Paul.
Vaughan, Diane. 1996. The Challenger Launch Decision. Chicago: University of Chicago Press.
Vaughan, Diane. 1999. The Role of the Organization in the Production of Techno-scientific
Knowledge. Social Studies of Science 29 (6):913–943.
Waldrop, M. Mitchell. 2001. The Dream Machine: J. C. Licklider and the Revolution That Made
Computing Personal. New York: Viking.
Waldrop, M. Mitchell. 2003. Autonomic Computing: An Overview of the Concept and Exploration of
the Public Policy Implications. Washington, DC: Woodrow Wilson International Center for
Scholars.
Wegner, Daniel M. 1995. A Computer Model of Human Transactive Memory. Social Cognition 13
(3):319–339.
Werhane, Patricia H. 1999. Moral Imagination and Management Decision Making. Oxford: Oxford
University Press.
Werhane, Patricia H., and Michael E. Gorman. 2005. Intellectual Property-Rights, Moral Imagina-
tion, and Access to Life-Enhancing Drugs. Business Ethics Quarterly 15 (4):595–614.
Westrum, R., and H. A. Wilcox. 1989. Sidewinder. Invention and Technology (Fall): 57–63.
Winner, Langdon. 1986. The Whale and the Reactor: A Search for Limits in an Age of High Technol-
ogy. Chicago: University of Chicago Press.
6 From Wizards to Trading Zones: Crossing the Chasm of
Computers in Scientific Collaboration
Jeff Shrager
The Chasm
Scientists are the prophets of the modern age. Whereas in the past prophets repre-
sented God to the masses, today scientists represent reality. As consumers, we are all
quite used to computers as our interface to reality—your phone conversations are
transmitted by computers, your bank account is virtual money, and nearly everything
that you see or hear in the media is refracted through computers. The same is true,
and even more so, in science: scientists don’t fly around in space sketching pictures
of the earth’s weather, nor do they peer into atoms with microscopes. These are the
jobs of instruments whose outputs are nearly always computer-mediated. But whereas
consumers don’t worry much about being held at arm’s length from reality by com-
puters (perhaps we should worry more), scientists worry about it quite a lot. Although
we have nominated scientists as our interpreters of reality, it’s actually not scientists
but engineers, and more and more often software engineers, who have direct access
to reality. Software engineers stand between scientists and the instruments that really
read the photons. Thus scientists are left to interpret the “massaged reality” offered
up by software engineers. But software engineers are essentially carpenters with no
special training in science, and so the relationship is problematic.1 No prophet would
last long if, rather than claiming a direct line to God, he was just repeating what some
carpenter told him!
I’ve spent nearly three decades working to bridge this chasm between scientists and
reality, most recently between molecular biologists and computer scientists—to
arrange it so that molecular biologists don’t need to have reality mediated by software
engineers, or at least so that they have a deep enough understanding of computing
that they don’t worry about this arrangement. Some of my early attempts were, in
retrospect, misguided. However, I believe that they are worth describing for two
reasons. First, there are still many similar projects being pursued today; perhaps the
108 Jeff Shrager
present analysis can help those researchers avoid plummeting into the chasm. More
directly, the history of my (mostly misguided) ideas eventually did lead me to what I
think is actually a very good solution to the problem of software engineers mediating
science. Of course, I could be wrong again; this new idea could be yet another mis-
guided one. Time will tell.
Before getting into my (mostly failed) attempts at bridge building, let’s examine a
few examples of the breadth and depth of the chasm that we need to cross.
The history of statistics in biology is quite interesting. Biology provided many
fundamental examples in the development of modern statistics. But with the advent
of molecular biology, biologists’ ability to access the discrete components of life (DNA
and such) has led them to feel that they could use simple tools like microscopes, gels,
and imaging to access the machinery of life, making statistics irrelevant. As a result,
molecular biologists over the past three decades were almost never trained in serious
statistics. Unfortunately, and inevitably, as the complexity of our understanding of
life mounts, sensitive instrumentation is taking over the field; one can no longer make
much use of the view through the microscope lens to understand what’s going on in
a cell; instead one needs laser spectrometers, microarrays, quantitative PCR, pathway
analysis, multiple testing corrections, dynamical systems models, Bayes nets, and other
such machinery that require quite subtle computation and statistics. Although biolo-
gists are becoming increasingly aware that they need to undertake statistical analyses,
they have little idea how to do them, and so they rely upon statisticians (or, more
often, students who could only play statisticians on YouTube), taking these folks’ word
for the results. This gives these (sometimes pretend) statisticians enormous say and
sway over what gets done in a lab, because statistics demands a certain amount of
data to get significant results. Unfortunately, the lab side of biology requires enormous
time and labor, and getting even small amounts of data can cost a great deal. More
than once, biologists have come to me with some data that they spent (not unusually)
a year and (not unusually) hundreds of thousands of dollars obtaining, only to find
out, after I do just a few minutes of math, that that they would need many times that
amount of data to get any useful results, because of variability in the data. The upshot
of this sort of interaction is almost never pretty; the biologists often end up putting
the paper in journals with less and less prestige until they manage to get it through
peer reviewers who also don’t understand statistics. This problem is slowly being dealt
with through new kinds of Bayesian statistical tools that incorporate biologists’ subjec-
tive priors, and biologists are (re)learning statistics. The most important positive trend,
which foreshadows where I’m going in this chapter, is that biologists are starting to
design experiments in close collaboration with statisticians and computer scientists.
Put another way, biologists are becoming willing to include statisticians and computer
From Wizards to Trading Zones 109
scientists as first-class members of their research collaborations, as opposed to treating

them like hired help and expecting them, as an afterthought, to “find the results” in
the data.
Another disconnect between biologists and computer scientists, implicit in the
above example, has to do with the speed with which results can be obtained. I
described the way that biologists will regularly spend years and hundreds of thousands
of dollars producing data which the statisticians proceed to demolish in minutes (or
at most, a few days) sitting at a computer, essentially for free. The opposite is also
true, and in some ways more problematic: computer scientists can spend the same
few days at their computer (essentially for free) and obtain new results by reanalyzing
the biologists’ own data, after (or sometimes even before) it is published! Imagine that
you are a biologist who has sweated blood, perhaps spending your whole dissertation
on producing a small pile of data, which your lab has poured perhaps a million dollars
into. You publish your interpretation of these data, and (as is required) make the data
public.2 A week later you find that some computer scientist has downloaded your data,
and has not only demonstrated that your interpretation was wrong, but has produced
a new, more correct conclusion based upon computational analyses of your own data.
If the computer scientist is feeling particularly magnanimous, he’ll ask you to coauthor
a paper, but there’s no ethical requirement for him to do that; your data are public—
all he ethically need do is to cite you. This may feel to you like intellectual theft, but
it’s no more intellectual theft than it is to use E = mc2 in a calculation, referencing
Einstein (which no one does, but it’s obvious where that came from). I have, more
than once, found myself on either the short or the long end of this stick, or worse,
straddling both sides trying to keep scientists from having one another hauled up in
front of ethics boards for just this sort of thing. Not only does it happen all the time
(not so often involving ethics boards, thank goodness), but it will be happening more
often as more and more data become available and as computers become more and
more powerful, able to mine larger and larger quantities of data. Unless biologists get
to the point where they can mine their own and others’ data in this way, or else form
early and strong collaborations with statisticians and computer scientists, the biolo-
gists are the ones who will end up as “mere” worker bees, gathering data, while the
computer scientists will be the producers of theoretically important results.
Into the Chasm
So this is the chasm that we want to find our way across: biologists need to be able
to deeply understand computation and, to a large extent, to do it themselves (or so
it seemed to me). Before diving into a survey of my failed attempts to cross this chasm,
110 Jeff Shrager
I’d like to clarify that I will mostly discuss molecular biology here because that’s the
field through which I have come to this problem. The problem is not the same for all
fields. Many scientists learn to program, and do so all the time. For example, the whole
concept of computer simulation was developed in large part by and for physicists, and
many physicists are highly competent programmers. Unfortunately, biologists don’t
have such a tradition, and, as we shall see, computation hit biology hard at the wrong
time, and has nearly overwhelmed the field. Although biocomputing is ahead of other
fields in knowledge representation (what is sometimes called “semantics” or “ontol-
ogy”), biological knowledge bases are rather haphazard affairs, connected to one
another in an ad hoc manner. In order to do computing in this space, one needs to
have a strong stomach for intense kludgery.3
But the same biologists who kludge things together in their labs all the time don’t
seem to have the stomach for it in computers. Why do physicists? Perhaps it’s because
mathematics—the knowledge representation of physics—isn’t as bad a kludge as is
most biological knowledge representation? Maybe it’s that computers, not being alive,
simply don’t interest biologists? Maybe it’s because biologists have traditionally been
able to get along without them? Another reason comes from an interesting historical
accident: biologists began finding it impossible to live without computation around
the early 1990s. This was coincidentally about the same time that computers had
begun to appear simple, by virtue of the Web’s emphasis on simple pushbutton func-
tionality. As a result of this accident of historical timing, biologists came to see soft-
ware in the model of a vending machine: you put in a quarter (some data), press a
button, and out pop results. I’ve often encountered the attitude among biologists that
“There must be a Web page somewhere that does this. … I’ll just surf around until I
find it, and then put my data in and press the ‘compute’ button, and results will pop
out.” This is generally true for consumer needs—if you want to buy a book or schedule
a flight—but not so much if you’re doing science. And even if you do get lucky and
find that someone has thought about the same problem before, and has left you a
way to do it, it’s usually nontrivial to get it working, and in any case you’d better
understand what it’s doing or you’re likely to misinterpret the results. These are the
capabilities that are expressly missing in biology, and these are what lead to the chasm
that I’m trying to cross.
Wizard: Computers as Collaborators
The buttress of my attempts to build bridges is my master’s degree project, advised by

Tim Finin at the University of Pennsylvania. As an undergraduate computer scientist,
I developed help systems for the Penn engineering community’s large time-sharing
systems. In my master’s work, Tim Finin and I went a step farther, developing an AI-
based assistant for computer users, called Wizard (Shrager and Finin 1982). Although
I doubt that I realized it at the time, our idea had significant subtlety, and it directly
underlies the rest of my work. Rather than just being another help system, Wizard
fundamentally changed the role that help systems played with respect to their users:
instead of being mere tools, the help system could become a collaborator. What is the
difference between a tool and a collaborator? Tools don’t understand the goals of the
project and don’t take initiative to reach those goals. Collaborators, on the other hand,
do understand and share the goals of the project, and by that virtue can take initiative
to reach them. Humans wield tools, but do not wield their collaborators! Instead, we
work with them; the project becomes a shared one, with shared goals and negotiation
about who’s going to do what. You don’t negotiate with your tools. With this in mind,
Tim and I focused on goals, and specifically on goal recognition—that is, giving the
computer enough of an understanding of the user’s goals that it could recognize them
from the user’s actions, and take initiative to help. Out of this arose Wizard, which
was probably the world’s first computer program that was able to take the initiative
in offering assistance to a user (Shrager and Finin 1982).4
Now, as a young computer scientist working in AI, I believed that psychologists
knew something that I didn’t—specifically, that they understood things like goals and
how goal recognition could work. I had taken some psychology courses as an under-
graduate and decided, somewhat insanely, that I didn’t just want to study psychology,
I wanted to actually be a psychologist. (Watch as this irrational need to go native
reappears as a theme in my life!) Not wanting to leave computing too far behind, I
undertook doctoral work at Carnegie Mellon University (CMU), where there is a strong
computational cognitive psychology program.
At CMU I pursued several apparently unrelated topics: computational modeling of
cognitive development (with David Klahr and Bob Siegler), computational modeling
of adult cognition (with John Anderson), and computational modeling of scientific
discovery (with Pat Langley and Herb Simon). I shall leave for another time the inter-
esting analysis of how computer wizards, cognitive development, and scientific dis-
covery relate. Suffice it to say that goal recognition is an observational discovery task
(Cohen, Perrault, and Allen 1982), analogous to scientific discovery, and that one
important way to obtain an understanding of how goals work in a given culture is to
have grown up (developed) in that culture (Collins, Clark, and Shrager 2008).
My first real foray in bridging the chasm between biologists and computation lay
at the confluence of the concepts of “computers as collaborators” and “computers as
112 Jeff Shrager
scientists.” Whereas it seemed to me impossible at that time (and still does today) to
build an artificial scientist, I thought that perhaps we could build computational tools
that could participate with scientists as true collaborators, at least at the level of
wizards like Wizard. Such tools could have a sense of the goals and, even more impor-
tantly, could provide explanations. Explanation was a concept that we had only a
glimmer of in Wizard, but it is obviously an important part of science (e.g., Hempel
1965). In the current context of collaboration: you don’t explain to your power tools
what you’re doing with them, and you don’t expect explanations from them when
they make mistakes, whereas you do explain things to, and expect explanations from,
your collaborators.
I thought that by endowing the program with an ability to understand scientific
goals (even in a simple sense), and to understand and provide scientific explanations
(even in a simple sense), scientists would be that much closer to not having to depend
upon software engineers (even in a simple sense), because they would be able to col-
laborate with the program (even in a simple sense). To be somewhat less fantastical,
one can think of explanations on the input side as a sort of programming language,
and on the output side as a sort of debugging log. These would be rendered into
“plain” language—or, in this case, scientific language, that is language that scientists
can understand without having to be software engineers. If the scientists could interact
with the program’s logs in the scientists’ own terms, and perhaps even write programs
in the same terms, there would (I believed) be less need for scientists to engage software
engineers to speak in computational tongues.
Afferent, a Computational Scientific Collaborator: A First Run at the Chasm
At this time—around 1996—another AI engineer, David Chapman, asked me to help

him productize a robotic drug-discovery tool, called Afferent, that he had prototyped
while working at a biotech. Afferent and its relationship to drug discovery is described
elsewhere (Shrager 2001). Here I highlight the ways in which Afferent could be
thought of as a collaborator, in that it could be said to understand the medicinal
chemists’ goals and to offer explanations.
One of the most important approaches to discovering new drug leads is called “high
throughput combinatorial chemistry” (combi-chem). Combi-chem works by running
numerous chemical reactions, thereby creating numerous molecules that could become
drugs. Now, chemists realize that running numerous reactions is a robotic problem
(at least for large values of “numerous”), but chemists are too busy being chemists to
learn to be robot programmers as well. Chapman was a robot programmer, and I
Figure 6.1
A chemical reaction as entered into Afferent. Note its similarity to the notation used by chemists,
such as that found in a textbook.
worked in AI, human-machine interaction, and computational models of science. We,

along with several other engineers, produced Afferent, a program that would allow
medicinal chemists to conveniently get a robot to do combi-chem.
Afferent approached this task through a combination of subtle technologies woven
together through a user interface that walked a difficult line between power tool and
collaborator. Afferent’s interface enabled the chemists to “talk chemistry” to the robot.
The chemists entered chemical reactions and protocols into Afferent in more or less
the same format that the chemists were used to—the sort of thing that you’d see in
a chemistry textbook (figure 6.1).
Given reactions and execution protocols for the chemistry, Afferent would operate
the robot to run the reactions. At the same time, Afferent would simulate the chem-
istry. There are two reasons for bothering with this very complex simulation step. First,
the chemists want to be able to tell what products to expect from the reactions. More-
over, chemistry is finicky; many (sometimes most) of the desired reactions won’t work
as predicted, and you want to be able to tell which ones failed. This is done by various
analytical measures that compute the amount of product in each resulting reaction
vessel. Failed reactions can be traced through the reaction simulations to locate offend-
ing reaction pathways. This sort of analysis and the corresponding traces constitute
one important type of explanation in chemistry. (See Shrager 2001 for more details.)
Whereas Afferent could be said to “read” chemistry, “think” (simulate) chemistry,
and “speak” chemistry, it wasn’t anything like a contributory expert in chemistry. It
114 Jeff Shrager
contributed nothing in terms of guiding heuristic search through reaction space.5

Regardless, Afferent went far beyond a simple combi-chem tool, and could fairly be
said to be, in some sense, a collaborator in the drug discovery process.
Successful as Afferent was in combi-chem, the goal of a true computational scien-
tific collaborator remains elusive. No computer program can (yet) rise to the level of
subtlety required of a true collaborator. Collins and Kusch (1998) argue on theoretical
grounds that computers will never be able to rise to this level (but see Collins, Clark,
and Shrager 2008). There is also an issue of “explanatory trust” that cannot be bridged
through “mere” AI: although human software engineers might misunderstand specs,
make mistakes, and miss milestones, at least you can have a conversation with them.
They can explain what they are doing or why they are not succeeding, and one can
bring managerial or social pressure to bear on them. It’s impossible to bring social
pressure on a computer program. Nonetheless, Afferent was perfectly good as a power
tool, and sold well. In 2000 the company was sold to MDL, Inc., and I made enough
money to fund my next attempt at crossing the chasm.
Becoming a Molecular Biologist: Trying to Leap the Chasm
Software engineering and psychology stand apart from most other technical fields in
that they are, in an interesting way, “domain independent.” When you write a
program, the program has to be computing about something else. It is similar in psy-
chology: when you study someone’s thinking, the person you are studying has to be
thinking about something else. This makes computation and psychology quite flexible,
in that, as a practitioner of these fields, one can work on whatever in the world one
likes. Suppose, for example, I wanted to live on top of a mountain and work in
astronomy; all I’d have to do, as a software engineer, is hire on to write programs for
astronomers. Similarly, as a psychologist of science, I could decide that how astrono-
mers think is somehow particularly interesting, and that I want to study them. Many
other fields that are together called the “social sciences” share this property, but not
so most technical fields, which are usually about just what they are about, not what
some other field is about. So in being a computational psychologist of science, I’m
able to study whatever I like in the scientific world. And because of the critical role
that software engineers play in science, I could potentially make significant (if silent)
contributions to it.
Now, there is a strong tendency for practitioners of metafields to get pulled into
the target field—to begin seeing the world as the community under analysis sees it.
After all, that community must be pretty interesting to you; otherwise, why would
you be studying it? This tendency has long been recognized in anthropology, where
they call it “going native.” To “go native” as a scientist means to begin finding the
domain of the science more interesting than the metascience, and to adopt the
scientists’ explanations for their own actions. In the limit, going native in science
might mean, in Collins and Evans’s (2007) terms, becoming a “contributory expert”—
becoming a real astronomer, rather than merely working as a programmer for astrono-
mers, or as a psychologist studying astronomers. I felt this tendency quite strongly
with biology, and for a time even went native myself.
After Afferent was sold, I made up my mind to leave computers behind and do
“real” science. To this end, I stopped programming altogether and set out to become
a “real” molecular biologist. The fact that computers had become essential at that time,
yet also essentially mysterious to biologists, offered me a way into the lab of my choice.
I made a deal with the biologists: you teach me to be a molecular biologist, and I’ll
do your computing for you. Who could refuse?
In addition to satisfying my romantic need to go native in biology, becoming
a molecular biologist myself was, in my fantastical imagination, another route for
scientists to understand computers. In this fantasy, no one would begin in biology
per se; instead, one would always undergo a dual training: first you’d become a com-
puter scientist, and only then would you study biology. Under this plan you become
a contributory expert in biology, but with a strong background in computing. These
disciplines could be undertaken nearly simultaneously; and indeed they now are—for
example, in specialized biocomputing degree programs. This is not a brand-new idea:
Seymour Papert (1993) had the idea that everyone should learn to program in elemen-
tary school, and he invented a simplified programming language, called Logo, that
supposedly made programming easy and motivating for kids. But Papert theorized
that programming somehow prepared one better for general thinking. My idea was
much more specific and practical: all scientists needed to be programmers in order to
cross the chasm between themselves and reality—ground now held fast by software
engineers. In my model everyone is first a software engineer, only later applying com-
puters in biology. You end up as a contributory expert in biology, but you arrive with
a strong background in computing.
Of course, there’s nothing particularly new about the idea of learning the important
background (computing or math or whatever) and then taking on a specific domain.
It was just apparent to the biologists, and to me, that biology was becoming a much
more computational and mathematical field that current biologists were unprepared
116 Jeff Shrager
Figure 6.2
The author engaged in some mysterious bench molecular biology. Note the tools of the trade:
multiple timers (hanging), pipettes, various bottles, safety glasses, notebooks. (See Shrager 2004
for more detail.)
for—thus the chasm. Because working biologists have no time to retrain in computing,
and because computation has the flexibility of needing a domain, computer scientists
could be turned into biologists, thus bridging the chasm.
The fastest way to perform this experiment was, of course, to inject the drug myself
(figure 6.2). This was, for me, in large part, a successful experiment: I coauthored
numerous almost purely biological papers (e.g., Labiosa et al. 2006), and was co-
principal investigator of several almost purely biological grants.
Unfortunately, once in the lab, no matter how hard I ran from computers, I was
unable to hide from them, and programming came to occupy more and more of
my time. Not only did I have my own biological computing to do, but as none
of the other “real” biologists knew much about computers, I was servicing their ever-
increasing computing needs as well. (Of course, this was the explicit deal I had
made with them in exchange for their teaching me molecular biology.) I was also,
as it turned out, pretty poor as a bench biologist. (I can’t cook very well either
and don’t enjoy it, but unfortunately, cooking is closest to the operational skills
utilized in a lab.) As a result, I was much more valuable to the lab as a computer
scientist than as a molecular biologist. So, after a couple of years of actual pipet-
ting, as shown in figure 6.2, I found myself working almost all the time at a
computer rather than at a bench. The inverse was true of the other biologists in
the lab; they were about as good at computing as I was at pipetting, so I ended
up with a long line of biologists at my back in need of computational help. My
remaining strong romantic desire to be a real molecular biologist, rather than a
software engineer working for biologists, led directly to the my next approach to
the chasm’s edge.
BioBike: You Can Drag Biologists to Computers, but You Can’t Make Them Hack
There was so much computing to do, even in my small lab of about fifteen people,
that it was beyond my abilities to keep up with. And I still harbored the romantic
fantasy of doing biology, not programming for biologists. In fact it was this tension—
my trying to pull toward being a molecular biologist, and the molecular biologists
trying to pull me into doing their computing for them—that led me to realize that
the chasm existed in the first place, and also to the next idea of a way across it. If
computers can’t themselves be scientific collaborators, and programmers can’t effec-
tively become biologists, maybe it would work the other way around: enable biologists
to do their own computing! Again, there is precedent for this in physics and other
fields. The language of physics is math, and programming math is what programming
was for in the first place; but what is the language of biology?
As I mentioned above, biological representation and computing were (and remain)
quite a mess. You could put a full-time software engineer in every lab to help biolo-
gists apply computers to their specific work, and still not be able to satisfy this ever-
increasing need. If biologists were to be enabled to do their own computing, one would
need to reduce many time costs: the cost of gathering the relevant data and knowledge
and wiring it all together, the cost both of learning to program and of learning the
numerous representational and programming languages needed to get even simple
things done in biocomputing, among other costs.
Onerous as these obstacles may seem, my colleagues and I elegantly addressed
all these needs, to some extent, through a novel Web-based technology called a
KnowOS—a Knowledge Operating System (Travers, Massar, and Shrager 2005), on
which we developed a specialized Web-based biological programming environment
called BioBike—a sort of Logo for biologists (figures 6.3 and 6.4; Massar et al. 2005).
118 Jeff Shrager
Figure 6.3
BioBike screens depicting (left, partly occluded) the built-in knowledge framework and (right,
overlapping) the command interface. The program depicted (right bottom, beginning “defun
...”) is all that is needed to compute molecular phylogeny among organisms based upon func-
tional orthology. (See Massar et al. 2005 and Travers, Massar, and Shrager 2005 for more detail.)
Figure 6.4
The BioBike/BBL visual programming interface. The computation depicted here (center) was
created interactively through the Web. It computes the mean length of the coding genes of the
cyanobacterium Synechocystis sp. PCC 6803. (Results are at bottom.) Contrast this with the messy
complexity of figure 6.3.
BioBike offered unprecedented integration capabilities and made biocomputing as

easy as we were able to imagine. I taught a course at Stanford in using BioBike, and
my colleague Jeff Elhai, a biologist at Virginia Commonwealth University, taught it
there. Elhai continues to work to make BioBike even more biologist-friendly (figure
6.4). He even sent a postdoctoral biologist, Arnaud Taton, on a world tour to train
biologists to use BioBike in situ, helping them conduct their own research in their own
labs. Still, to date, significant numbers of biologists have not adopted BioBike or any
other programming paradigm.
Let me give you just one example of how difficult this chasm is to cross, and to
understand. Once BioBike was up and running, I began to introduce it to the long
line of biologists wanting me to do computing for them. I started by giving lectures
in order to acquaint them with the general concept of biocomputing, and specifically
how to use BioBike to do real work. These were well attended by enthusiastic biolo-
gists. When I worked on a particular project for a biologist, I used BioBike to build a
120 Jeff Shrager
small quick-and-dirty application for them; all that they usually had to do was to make
small modifications to the code to get their work done. Figure 6.3 depicts such a
program, which does molecular functional phylogeny. This was as close to the scien-
tific vending machine as one could imagine; it was even better, in fact, because the
functionalities were tuned to the particular task at hand. I even wrote detailed instruc-
tions, in lablike protocol format, explaining how to use these programs.
Nevertheless, I discovered that you can drag biologists to computers, but you can’t
make them program! I once created a three-line BioBike program for a biologist in our
lab. If he had used my code, his computation would have taken about ten minutes.
Several weeks later he came back to me reporting that he’d finally found the time to
accomplish the computation. The time? Yes, he’d spent all weekend on it! It turns out
that he had used BioBike only to download the data into a spreadsheet. (He had actu-
ally given me the data in a spreadsheet in the first place, and I had simply uploaded
it into BioBike, so why he had to re-download it into a spreadsheet is entirely beyond
me.) Next he had arduously applied spreadsheet functions (sorting, deleting, and
simple math) to get the desired result. My program (which I’d given him detailed
instructions on how to use) would have done all of that work for him automatically,
taking ten minutes to get the same result that he had probably spent twenty hours
on.
This was not an isolated case; this sort of thing happened to me over and over.
A biologist would come to me to ask for help; I would try to explain BioBike and create
a very simple example program that she or he could modify and use almost immedi-
ately. It would have taken perhaps an hour of preliminary attention to learn how to
use BioBike, and then perhaps another hour (and these are high estimates!) to get the
work done. Not once that I can recall in my five-plus years of trying to get this to
work did a biologist ever take the short time it would have taken to save himself or
herself hours, or days. They would either walk away saying that they would do it, and
then wouldn’t; or would walk away confused; or would just walk away.6 Most com-
monly they would simply keep annoying me about it until I did the work myself. In
this last case BioBike saved me enormous time; it was, and remains, an excellent bio-
computing platform—saving biocomputing engineers from the most painful and
repetitive parts of their work—and in this sense it was (and continues to be) successful.
But as far as biologists doing their own computing, so far BioBike has, in my experi-
ence, failed to live up to its promise.
Although I have given up on that goal, Jeff Elhai and his coworkers have persisted,
taking BioBike even farther toward the Logo vision of biocomputing by biologists.
They have created a simplified push-button, domain-specific Web tool for genomic
computing, including a special-purpose, simplified programming language, called BBL

(Massar et al. 2005), and they have developed an impressive visual programming
interface for BBL, depicted in figure 6.4.
BioBike as Trading Zone: Collaborating without Crossing the Chasm
I failed in my effort to create a computational collaborator. I failed in my effort to

exemplify the transition between computer scientist and biologist. I failed in my effort
to train biologists to program, even after providing them with a platform that made
programming simple. But from all these failures has arisen a potentially important,
interesting, and entirely unexpected idea.
Although biologists did not themselves end up programming in BioBike, a com-
munity arose which does, in the end, accomplish the desired goals. Importantly, this
community does not consist of biologists who are autonomously programming their
own solutions. Rather, because BioBike is a Web-based platform, the biologists are able
to interact in a very tight collaboration with programmers directly through BioBike,
even if they are all remote from one another. In some cases this collaboration has
been effectively elbow to elbow: biologists and programmers sitting in different loca-
tions—often thousands of miles apart—working on the same problem through shared
BioBike screens. In this BioBike “trading zone” (borrowing a term first coined in a
scientific context by Peter Galison [1997]), real problems are really getting solved, not
by AI-based computational collaborators, not by the programmers becoming biologists,
and not by the biologists becoming programmers, but by the enabling artifact of
BioBike, a biocomputing-specific, knowledge-rich, Web-based computational trading
zone (Shrager 2007).
The specific role played by BioBike in this trading zone is complex and interesting.
On the one hand, it can be thought of as a sort of platform for the development of
what Star and Griesemer (1989) have called “boundary objects,” forming a shared
focus for the communities of practice that meet in the trading zone. But a program
is more than an object; it carries in and of itself a semantics, and so BioBike is more
like a dynamic book—a dictionary, perhaps, of the creole that forms in the trading
zone—than like a static object. Discourse in the trading zone runs through BioBike
rather than, or in addition to, running around (about) it. There are many features that
enable BioBike to serve this role as locus of a scientific trading zone, and I have dis-
cussed these in some detail elsewhere (Shrager 2007). The key point is that I neither
foresaw this result nor made an explicit effort to reach it. Instead, the BioBike trading
zone emerged from the activities of the community of BioBike users and developers.
122 Jeff Shrager
So I am left, now, with this last way across the chasm between biologists and
computing: scientists and engineers interacting as peers through a shared science-
specific collaborative computing platform. With twenty-twenty hindsight, it is clear
to me that this was the correct solution all along; the biologists needed to think of
the computer scientists as collaborators, not as power tools (or the operators of power
tools). In a strong sense, BioBike is not what is enabling this, so much as the his-
torical fact of the computer scientists’ not being under the thumb of the biologists.
As long as I was in the lab, working for the biologists, I was never going to be more
than a carpenter to them. But by enabling scientists—whether biological scientists
or computational scientists—who are at a similar level of contributory expertise in
their own fields to work with one another, not for one another, to cooperate—no,
to collaborate!—the needs of shared goals, explanation, and getting things done are
all met.
Or perhaps not. Time will tell.
Appendix: Popular Press regarding Shrager and Finin 1982
Each of the following was the very first sentence of popular news reports that arose
from Shrager and Finin 1982. The entire articles can be read at http://nostoc.stanford
.edu/jeff/personal/vita/pubs/1982WizardPress.pdf:
H. W. Pierce, “Computer Advises without Being Asked,” Pittsburgh Post-Gazette, August

16, 1982, 6: “Most people don’t like back-seat drivers. But what if the back-seat driver
were a computer?”
T. Hritz, “Busybody Computer,” Pittsburgh Post-Gazette, August 19, 1982, 4: “When I
read earlier this week about the National Conference on Artificial Intelligence which
is meeting here, and about the computer out at Carnegie Mellon University that gives
advice without being asked for it, I found it hard to believe. I always thought that you
couldn’t get anything out of a computer unless you put that thing into it.”
T. Henkel, “Digital Big Brother: Sassy Program Keeps Programmers in Line,” Computer-
world, September 6, 1982, 16: “Thanks to the marvels of technology, computers can
now be taught to torment their users.”
Notes
Mike Gorman edited in detail several drafts of this chapter, helping to hone the concepts and
language, but those are the least of his contributions! This paper arose from a talk presented at
the first symposium on Trading Zones and Interactional Expertise, hosted by Arizona State
University. I came to that meeting with a completely different—and much more shallow
and boring—talk in hand. After sitting through the first day of talks, especially those of Mike,
Peter Galison, and Harry Collins, I completely rewrote my talk, calling it (originally) “My History
of Dumb Ideas.” After hearing the talk, Mike hounded me to put it on paper for the present
volume, and helped me to fine-tune my thinking and to connect the concepts better to the
themes of the symposium. As a result, Mike has fundamentally helped me to understand my
own intellectual history.
1. This issue lies at the heart of Galison’s 1997 Image and Logic: “To get at the shifting meaning
of ‘the experiment and the experimenter,’ we will need to empathize with the anxiety that
physicists have often felt at their loss of control over the rapidly expanding laboratory, at the
complex confrontation of physicist with engineer” (p. 5).
2. Although it must be made public (especially when the project is publicly funded), data is often
not actually made easily available. In the best case, it is just downloadable, but I have many
times found that I have to ask multiple times for the data, and then get it in dribs and drabs, or
in completely uninterruptible formats.
3. A “kludge” (the u is pronounced as oo in “boot”) is a term of art in computer science, referring

to a usually messy, ad hoc, and poorly constructed solution to a complex problem. A kludge gets
the job done, but because it is not a good solution, it is usually hard to understand, modify, and
support.
4. The reader will be most familiar with the wizard concept in the guise of the much-maligned
Microsoft (MS) Paperclip, which appeared about a decade after Finin and I published our work.
Although the MS Paperclip did essentially the same job as Wizard, what people didn’t like about
the Paperclip seems to have been its arrogance and annoying animatronics. The Paperclip was
retired around the turn of the century. It was pretty clear, even to nontechnical observers, that
the concept of a computer that could take its own initiative was going to be problematic. In the
appendix, I reproduce some responses to our work that appeared in the popular press at the time
of its publication.
5. We did have in mind trying to do this, but never got around to it. And, in retrospect, we
probably wouldn’t have got very far.
6. Jack Carroll and Mary Beth Rosson refer to this as the “production paradox” in their eloquent
and classic paper “The Paradox of the Active User” (Carroll and Rosson 1987).
References
Carroll, J. M., and M. B. Rosson. 1987. Paradox of the Active User. In Interfacing Thought: Cognitive
Aspects of Human-Computer Interaction, ed. J. M. Carroll. Cambridge, MA: MIT Press.
Cohen, P. R., C. R. Perrault, and J. F. Allen. 1982. Beyond Question Answering. In Strategies for
Natural Language Processing, ed. W. Lehnert and M. Ringle, 245–274. Hillsdale, NJ: Lawrence
Erlbaum Associates.
124 Jeff Shrager
Collins, H. M., A. Clark, and J. Shrager. 2008. Keeping the Collectivity in Mind? Phenomenology
and the Cognitive Sciences 7:353–374.
Collins, H. M., and R. Evans. 2007. Rethinking Expertise. Chicago: University of Chicago
Press.
Collins, H. M., and M. Kusch. 1998. The Shape of Actions: What Humans and Machines Can Do.
Galison, P. 1997. Image and Logic. Chicago: University of Chicago Press.
Hempel, C. G. 1965. Aspects of Scientific Explanation and Other Essays in the Philosophy of Science.
New York: Free Press.
Labiosa, R., K. Arrigo, A. Grossman, T. E. Reddy, and J. Shrager. 2006. Examination of Diel
Changes in Global Transcript Accumulation in Synechocystis. Journal of Phycology 42
(3):622–636.
Massar, J. P., M. Travers, J. Elhai, and J. Shrager. 2005. BioLingua: A Programmable Knowledge
Environment for Biologists. Bioinformatics (Oxford) 21 (2):199–207.
Papert, S. 1993. Mindstorms: Children, Computers, and Powerful Ideas. New York: Basic Books.
Shrager, J. 2001. High Throughput Discovery: Search and Interpretation on the Path to New
Drugs. In Design for Science, ed. K. Crowley et al., 325–348. Hillsdale, NJ: Lawrence Erlbaum.
Shrager, J. 2004. On Being and Becoming a Molecular Biologist: Notes from the Diary of an Insane
Cell Mechanic. In New Directions in the Study of Science and Technology, ed. M. Gorman et al.
Mahwah, NJ: Lawrence Erlbaum.
Shrager, J. 2007. The Evolution of BioBike: Community Adaptation of a Biocomputing Platform.

Studies in History and Philosophy of Science 38:642–656.
Shrager, J., and T. Finin. 1982. An Expert System that Volunteers Advice. In Proceedings of the
Annual Conference of the American Association for Artificial Intelligence, 339–340. Menlo Park, CA:
AAAI Press.
Star, S. L., and J. R. Griesemer. 1989. Institutional Ecology, Translations and Boundary Objects:
Amateurs and Professionals in Berkeley’s Museum of Vertebrate Zoology, 1907–39. Social Studies
of Science 19:387–420.
Travers, M., J. P. Massar, and J. Shrager. 2005. The (Re)Birth of the Knowledge Operating System.
International Lisp Conference, Stanford University, June.
7 Authenticity, Earth Systems Engineering and Management, and
the Limits of Trading Zones in the Era of the Anthropogenic Earth
Brad Allenby
Introduction: The Anthropogenic Earth
Humans now live in a world that is fundamentally different from anything known
from earlier experience. It is a world where the critical dynamics of major earth
systems—whether they are predominantly atmospheric, biological, or physical, or, for
that matter, cultural, economic, or technological—increasingly bear the imprint of the
human. Indeed, the anthropogenic earth is characterized by large and complex inte-
grated human/natural/built systems, with complicated biological, physical, gover-
nance, ethical, scientific, technological, cultural, and religious dimensions and
uncertainties (Allenby 2005). As Nature put it in a 2003 editorial, “Welcome to the
Anthropocene”—roughly translated, welcome to the Age of Humans. This does not
mean “deliberately designed by” humans, for many aspects of these systems, although
human in origin, have not been consciously designed by anyone: the Internet, for
example, is clearly purely human in origin, and yet its complex information and
physical frameworks, like its broader social and cultural implications, have not been
designed by anyone; rather, they are emergent characteristics of this very complex
system. Nonetheless, it is apparent that a principal result of the Industrial Revolution
and associated changes in human demographics, cultures, technology, economic
systems, and belief systems has been a world increasingly dominated by human activ-
ity. This world is characterized by three major factors—accelerating change, complex-
ity, and radical contingency. Each is critical in its own right, and all interact with each
other in increasingly unpredictable but fundamental ways.
A few examples may help make the somewhat vague concept of the Anthropocene
a little more concrete. Start with what is perhaps the most fundamental physical aspect
of a planet, its characteristic radiation emissions spectrum. The earth’s spectrum,
however, is not just a matter of reflections from clouds, emitted infrared radiation,
and the like. Rather, it includes television and radio broadcasts, and leakage from all
126 Brad Allenby
sorts of technologies. The most accessible image of this emission is the well-known
picture of the earth from space at night, showing electric lights spread over North
America, Europe, and Asia (NASA 2008). In the Anthropocene, even the radiation
spectrum of our planet carries a human signature.
Similarly, almost everyone is aware of global climate change. Stand away from
the Kyoto Protocol—the international attempt to reduce climate-changing emissions
—and its surrounding hysterics, however, and a longer perspective is revealed: that
process, fitful and ad hoc as it is, represents the dawning of a realization that,
regardless of what happens in the short-term negotiations, the connection between
climate (and other natural systems), human economic and technology systems, cul-
tural and psychological patterns, and the built environment is so complete that
there are no solutions, only ongoing adaptation—there is no endpoint, no solution,
only dialog. This process is not an interaction among human communities, but a
continuing interaction within the human species, and between it and many other
natural and built systems that will continue so long as humans exist in numbers
anywhere close to our current population. The current negotiations surrounding the
Kyoto Protocol have been problematic in part because they involve only a few of
the affected communities, such as environmental and sustainability advocates, but
not others, such as those whose lives would be significantly affected by some of the
proposed responses. In some ways, the negotiations represent an enforced trading
zone and are unlikely to be successful, because widespread acceptance of the results
of the dialog is necessary (Collins, Evans, and Gorman 2007; and see chapter 2 of
this volume). Equally problematic, however, is that the Kyoto Protocol has been
approached as a stand-alone initiative, rather than as the beginning of a process
that, at this point, can have no foreseeable end. Policies can reduce—more likely,
redistribute—some of the human impacts on complicated and interrelated earth
systems, such as the climate system, but the growing human influence cannot be
eliminated (Allenby 2007a). Moreover, these particular perturbations are all part of
interconnected global systems, and a population of well over six billion humans,
each seeking a better life, ensures that the human role in global systems will only
continue to increase, absent some sort of population crash. The human relationship
with these global systems is not something that can be “solved”; rather, it is at this
point in our history an existential aspect of human existence (Allenby 2005). Thus,
even at the very beginning of our discussion, we can begin to see not just theoreti-
cal shortcomings in current governance models, but very pragmatic steps that can
be taken, in education and in practice, to respond to the challenges of the anthro-
pogenic earth.
Earth Systems Engineering and Management and the Limits of Trading Zones 127
In order to understand the implications of these systems for trading zones and,
indeed, for the engineering and management of earth systems in the future (Allenby
2007a), it is useful to take a closer look at two quite different types of system:
resource regimes and technology systems. Despite their obvious differences, four
commonalities will emerge. First, these earth systems all integrate human, natural,
and built components, albeit to different degrees—climate change is complex because
in large part it represents precisely such an integrated system, and thus engages
very different belief systems and ontologies. Second, earth systems are adaptive
complex systems and, because of the human component, tend to display a truly
destabilizing reflexivity, as the results of models, projections, policy initiatives,
changes in natural state, new technologies, and much else immediately become part
of the very system that gave rise to them, thereby changing those systems in
unpredictable ways. Third, the rate of change of these coupled systems is in many
cases not just increasing but accelerating. This appears to be true with global climate
change and the carbon and nitrogen cycles; it is certainly true of emerging tech-
nologies, discussed in more detail below. Acceleration is of particular importance
because the impact of changes in systems conditions depends not just on eventual
magnitude, but crucially on the speed with which changes occur. Ecologists, for
example, point out that the rate at which climate change occurs has far more impact
on the ability of species to adapt than does the eventual increase in temperature;
rapid changes are more destructive. Finally, and perhaps most subtle, earth systems
are interacting so as to create a global condition where virtually all assumptions
are increasingly contingent.
Resource Regimes
One well-known and important class of the integrated human/natural/built systems

that characterize the anthropogenic earth is composed of regional resource regimes
such as the Aral Sea or the Everglades—two examples that, in many ways, represent
opposite poles of this category. In the case of the Aral Sea, Soviet engineering and
management reflected state dominance of governance, specific aspects of Marxist
ideology and culture, and an economic, production-oriented focus. This approach
had many unfortunate environmental and social impacts. In contrast, a history of
centuries of exploitation of the Everglades is being countered by a policy based on
principles of adaptive management (Gunderson, Holling, and Light 1995) in hopes
of achieving a reasonable management regime that will reflect many different world-
views (NRC 2007).
128 Brad Allenby
The Aral Sea is actually a lake located in central Asia. At the beginning of the twen-
tieth century, the population in the area was approximately eight million, primarily
supported by irrigated agriculture. Irrigated agriculture in the region dates back to
approximately the sixth century and has always been a feature of human settlement
in the area. Over the last half-century, however, the Soviet Union diverted over 90
percent of the flow of the two feeder rivers, the Amu Dar’ya and the Syr Dar’ya, in
order to grow cotton, which was favored by Soviet central planners because it could
be exported for cash (UNEP-IFAS 2008). The United Nations Food and Agriculture
Organization estimated that historical inflow to the Aral Sea, some 56 cubic kilometers
per year, shrank to about 47 in the period 1966 to 1970, and had collapsed to 2 cubic
kilometers by 1985 (FAO 1998). As a result, the sea, which only decades ago was the
fourth largest lake in the world, has in a few short years lost about half its area and
some three-fourths of its volume, with concomitant drops in sea level and increases
in salinity. These changes took their toll on indigenous species. The number of bird
nesting species dropped from over 170 in earlier years to around 35 today. Increasing
salinity led to the loss of 20 out of 24 indigenous fish species. As always, these changes
were coupled with economic and social impacts: a vibrant local commercial fishery
saw a drop in the fish catch from 44,000 tons in the 1950s to zero today, with a loss
of 60,000 jobs, exacerbating the health impacts that toxic dust and degraded environ-
mental conditions were already imposing on local communities (ARAL 2003).
Substantial new areas of desert have been created in the Aral Sea region. By some
estimates, some 40 to 150 millions of tons of toxic dust per year are currently gener-
ated and spread primarily by the wind; blowing dust has also resulted in significant
melting of glaciers in the Pamir-Altay, Saili Alatau, and Akshirak areas. More funda-
mentally, perhaps, the substantial change in such a large geophysical feature has
modified regional climates, with potential impacts on the climate regimes of China,
India, and Southeastern Europe. In the region itself, the summers are shorter and drier,
and winters longer and colder, with reduced vegetative seasons and significant changes
in precipitation patterns (UNEP-IFAS 2008).
It is worth noting, however, that even after such extensive damage, mitigation is
possible; Kazakhstan is taking some restorative measures that appear to be having
some effect, while Uzbekistan is not (Pala 2006). Similarly, Iraq’s marshes, 18,000
square kilometers of which were destroyed in one generation by Saddam Hussein, were
devastated by 2000, but are being restored even while war continues in that country
(Lawler 2005). The Aral Sea is thus a lesson in what not to do, but such cases are also
indicative of what proper governance, earth systems engineering, and management
expertise can do even in difficult circumstances.
The Everglades, an area of approximately 18,000 square miles in southern Florida,

is also a water-based resource regime. Recent population growth has been robust. At
the time of the American Civil War, southern Florida was the most remote area in the
eastern United States and had fewer than 8,000 inhabitants; but as of 2002, some 6.5
million people were living in the area. Rock-mining operations for building homes
and attendant infrastructure have been an important, and politically powerful, local
industry with significant environmental impact. Despite its romantic image as an
ancient wilderness, the Everglades itself is only about five thousand years old, has
already gone through two very different regimes, and is embarking on a third (GAO
2003). The first was characterized by a broad and shallow flow south from Lake
Okeechobee to Florida Bay and the Gulf of Mexico; the second, a result of substantial
human modification beginning around the time of the Civil War and accelerating
strongly in the first half of the twentieth century, resulted in shifting drainage patterns
from Lake Okeechobee to the east and west, significantly reducing flows through the
Everglades; the third, reflecting a desire to reestablish substantial flows through the
Everglades, is currently evolving (GAO 2003; SUSF 2003). The technology that has
shaped the Everglades comprises not just the obvious (e.g., hydraulic engineering),
but the subtle and profound as well: for example, air conditioning for both homes
and cars was critical for attracting immigrants to the region, which experiences
uncomfortably hot and humid summers.
Over time, and reflecting industrial, agricultural, and demographic changes in
southern Florida, the Everglades has been altered by an 1,800-mile network of canals
built primarily over the last half-century, which has diverted some 1.7 billion gallons
of water flow each day. Equally important is the change in water quality that has
occurred. Originally the water of the Everglades was extremely low in the nutrient
phosphorous, and as phosphorous levels have risen, the dominant vegetation in the
Everglades has changed from saw grass, which requires low phosphorous levels, to
cattails. Subsequently, invasive species, especially melaleuca, Old World climbing
ferns, Brazilian pepper, and Australian pine, have been outcompeting native species,
aided by human disturbances such as draining marshes and more frequent fires. The
nesting success of birds, a predominant animal form in the Everglades, has declined
at least 95 percent since the mid-1930s (Kloor 2000). Changes in the Everglades have,
in turn, impacted Florida Bay, through perturbations in salinity levels, nitrogen
cycling, and other mechanisms, leading to die-offs in important species such as the
sea grasses (Schrope 2001).
In response to the imminent danger of the Everglades’ collapse as a functioning
ecosystem, a multibillion-dollar Everglades “restoration” project has begun (NRC
130 Brad Allenby
2007). Its intent is to restore water quality and water flow in nonhuman systems to
functional levels, while continuing to support industrial, agricultural, settlement, and
other human activity, and avoiding unacceptable flooding. The hope is to preserve
and rebuild a desired ecological community while maintaining human activity at a
politically acceptable level, a challenge that will require implementation of over 200
separate projects (GAO 2003). An important part of the plan is its emphasis on what
one might call a continuing dialog with both the human and the natural systems that
together compose the project, including continuing scientific review of both plans
and the natural systems’ reactions to what has already been done (Kaiser 2001). The
General Accounting Office report on the overall restoration effort explicitly assumes
that adaptive management is the approach being taken (although it also assumes that
this approach applies mainly to the project’s scientific coordination), but criticizes
how well it has been implemented (GAO 2003). Whether the conflicting goals of the
restoration project can be accomplished is unclear at this point (NRC 2007), but the
adaptive management approach, and the attempt to achieve agreement on design,
use, and management of the resource regime to meet multiple needs (including pro-
tection of ecological values), are an interesting and worthwhile experiment. The issue
also raises an interesting observation regarding the level at which trading zones
achieve rationality. It is likely that some of the participants hold conflicting mental
models of what the Everglades should be; for example, environmentalists and develop-
ers would desire different outcomes. Thus, they would not agree explicitly on any
particular outcome. Nonetheless, the dialog process could result in a socially efficient
outcome, in that explicit disagreements would be suborned by a compromise that
engaged sufficient support, and appeared transparent and democratic enough, to
achieve an acceptable consensus over some period of time. (This possible result would
be further complicated by the fact that mental models would no doubt be evolving
over that period as well—perhaps creating a form of “subversive trading zone” [Collins,
Evans, and Gorman 2007]).
One thing, however, is clear. The Everglades has long been, is now, and will con-
tinue to be for the foreseeable future, a product of human design and human choices.
As Schrope (2001, 128) comments, “Returning the Everglades to their natural state is
not an option. The project’s goal is instead to restore flow as much as possible while
also considering the needs of local residents and farms.” Like all complex systems, the
Everglades integrates its history into its future states; once perturbed, it does not return
to some hypothetical equilibrium. There is no “pristine” system to return to, and the
Everglades will never be “natural” again. It will be a designed and engineered system,
and it will display those characteristics that humans choose, including preservation
of flora and fauna, if that is a design objective and if the system can be engineered to
do so.
Neither the Everglades nor the Aral Sea is an atypical case, for water has been asso-
ciated with human settlement for thousands of years, and concomitant water manage-
ment activities have often led to unanticipated effects—soil salinization, susceptibility
to drought, impacts on aquatic and other species. Indeed, hydrological engineering at
all scales is not only archetypal but in many cases may well have been an important
contributor to the rise of civilizations, bureaucracies, and governance systems, as well
as engineering advances. In ancient China, for example, massive irrigation projects
were a sign of the development of highly organized feudal power as long ago as the
Chin state in the Wei valley, some 2300 years BP (Needham 1988, sections 1–7). Such
activities remain major sources of human impact on earth systems today. Examples
include the Pearl River delta in China, now undergoing massive and systemic develop-
ment; a $20 billion plan to replumb California’s rivers and deltas (which are already
extensively engineered); a similar effort to stabilize Louisiana’s wetlands, in light of
the damage done to New Orleans by Hurricane Katrina (somewhat less expensive, at
a projected $15 billion); Romanian plans to stabilize the Danube delta; and efforts to
revitalize the Baltic and Black seas (Schmidt 2001). It is not surprising, then, that the
Everglades reconstruction project has been visited by officials from Japan, Britain,
Brazil, Italy, Hungary, and other countries seeking to learn from it as they face the
need to undertake similar projects in their own jurisdictions.
That being said, the cultural dynamics that lead to projects like the Aral Sea irriga-
tion system are neither clear nor entirely rational. First, such projects reflect their
historical and political contexts, and thus they—and their impacts—are major exam-
ples of how the contingency and reflexivity of human systems are built into natural
systems. The Aral Sea project, for example, was affected by a number of historical
peculiarities, such as the Soviet fixation on economic development at all costs and the
underlying decision to exploit agriculture to create surpluses for military and industrial
development. Such policies obviously favored any projects that would facilitate the
production of export-oriented crops such as cotton. More fundamental and more
important, perhaps, was the Soviet Marxist predilection for demonstrating political
and even theological validity through large engineering works, as one might expect
from a society that took as its raison d’être the creation of a “new socialist man” at
the cost of millions of deaths, and undertook such massive projects as the Virgin Lands
initiative (Scott 1998, 96, comments on the “industrial-strength social engineering”
that characterized the Soviet Union’s version of Marxism). Such massive state projects
also reflect the desire for rapid economic development, which is a critical legitimating
132 Brad Allenby
factor for any state, and is especially important to the production-oriented, broadly
instrumentalist ethic of state Marxism (Harvey 1996; Shapiro 2001).
The Aral Sea is in many ways the opposite of the Everglades example, however. For
one thing, in the case of the Everglades, the current condition of the area was not
explicitly planned; rather, it evolved over time as the result of many intersecting, but
discrete, events and trends. The Aral Sea, in contrast, is an example of what can happen
as a result of the high modernist linkage of scientific and technological apparatus with
the political and cultural power of the state (Scott 1998; McNeill 2000). There were
no countervailing political forces opposing the technocratic, absolute state and its
belief structure. Moreover, unlike the Everglades project, the design intent of the Aral
Sea project was fixed: to build an irrigation system that would support cotton produc-
tion. There was no consideration of any ongoing dialog with the relevant human and
natural systems, nor were there plans to change behaviors or policy as needed due to
unanticipated changes in the system Rigidity, not resiliency, was thus built into the
system.
Technology Systems1
For most people, resource regimes and technology systems lie at opposite ends of their
psychological spectrum: one is natural and the other is artificial; one is associated with
environmentalism, sustainability, and the good, while the other is human, urban, and
problematic (Cronon 1995). This common characterization is dangerous, in part
because it hides the deep and growing integration of human, natural, and built
systems, often facilitated by technology systems. It thus enables a form of wistful
blindness. But it is dangerous not only because our technologies construct a human
earth. Not only are technologies more powerful than generally recognized, but an
entire integrated suite of technology systems are all changing at an accelerating rate.
As a result, relationships that have habitually and comfortably been regarded as stable
are so no longer; indeed, it increasingly appears likely that the human itself is being
converted into a design space.
The definition of economic periods in terms of technology clusters has a long
history: the “long wave” theory of economic growth identifies constellations such as
the beginning of the Industrial Revolution with water-powered textile mills, the
second “Kondratiev wave” of rail and steam, the third wave of steel and electrification,
and so on (Freeman and Louca 2001). This formulation and the scholarship underly-
ing it make a critical point that is too often overlooked: technology is an integrated
cultural process, not a collection of things (Bijker, Hughes, and Pinch 1997). For illus-
trative purposes, consider the historical example of a familiar—indeed, even faded—
technology system, the railroad. This technology did more than merely speed the
transit of goods or even change economic systems: in the mid 1800s, as it began its
rapid expansion phase, the railroad was not just the most impressive piece of machin-
ery most people ever saw; it was a sociocultural juggernaut that restructured the world.
Among the changes the railroads brought in their wake:
1. As a regional-scale integrated technology network, railroads required a uniform,

precise system of time, and thus created “industrial time” and its associated culture
(Rosenberg and Birdzell 1986). Before railroads, local times were isolated and charm-
ingly idiosyncratic: London time was four minutes ahead of Reading time, over seven
minutes ahead of Cirencester time, and fourteen minutes ahead of Bridgewater time
(Schivelbusch 1977). Interestingly enough, during the transition phase, time was first
made uniform within but not among railroad firms; stations serving several train
companies had different clocks (Buffalo had three different clocks at one point,
Pittsburgh six). Regional standard time did not gain legal recognition in the United
States until 1918 (Schivelbusch 1977).
2. Railroads created the need for, and coevolved with, national-scale communications
systems. In particular, telegraph technology was both coextensive with rail networks
(frequently laid along the same rights of way), and a necessary coordination mecha-
nism for the creation and operation of regional integrated rail systems (Grubler 1998).
3. Railroad firms shaped modern managerial capitalism. Whereas previous industrial
technologies had called forth a division of labor among factory workers, the scale of
railroad enterprises required a division of labor at the management level, with modern
and professional accounting, planning, human resources, and administrative systems
(Freeman and Louca 2001).
4. Railroad firms shaped modern capital and financial markets; railroad construction
was the single most important stimulus to industrial growth in western Europe by the
1840s. In contrast, the early factory system was supported initially by aristocrats and
landowners, and subsequently by factory owners using their own capital, an essentially
individualistic financial structure that was nowhere near adequate to support the huge
capital requirements of railroad firms (Freeman and Louca 2001).
5. Particularly in the United States, railroads became a potent symbol of national
power, and, more subtly, instantiated and validated the American integration of reli-
gion, morality, and technology. This was reflected in the language used by those who
134 Brad Allenby
viewed railroads as evidence of human ascension to godlike power (Daniel Webster,

for example), and those who viewed them as Satanic: “If God had designed that His
intelligent creatures should travel at the frightful speed of 15 miles an hour by steam,
He would have foretold it through His holy prophets. It is a device of Satan to lead
immortal souls down to Hell” (Ohio School Board, 1828, quoted in Nye 1994, 57).
Conversely, in the early 1800s the Western Railroad in Massachusetts urged ministers
to “take an early opportunity to deliver a Discourse on the Moral effect of Rail-Roads
in our wide extended country” (Nye 1994, 57). The railroads thus initiated a coupling
between religion and the technological sublime which forms an important basis of
American exceptionalism, and continues in the United States into modern times;
consider Robert Oppenheimer’s famous musing at Trinity Test Site at White Sands,
New Mexico, in 1945 at the first successful nuclear test, echoing Vishnu from the
Bhagavad Gita: “Now I am become Death, destroyer of worlds.” Notice also the critical
shift that technology enables: the aspect of destroyer of worlds in the Gita is divine;
at White Sands, the divine has become human (Allenby 2007b).
6. Railroads transformed landscapes at all scales, both physically and psychologically.
Chicago existed, and structured the Midwest economically, physically, and environ-
mentally, because of railroads (Cronon 1991). Psychologically, railroad technology not
only extended the sense of place and rhythm that previous transportation technolo-
gies, the horse and carriage and canal, had encouraged, but obliterated them. This
space-time compression is not, as some think, a purely postmodern phenomenon;
consider the psychological dislocation inherent in Heinrich Heine’s words in
1843 upon the opening of rail lines from Paris to Rouen and Orleans (quoted in
Schivelbusch 1977, 37):
What changes must now occur, in our way of looking at things, in our notions! Even the elemen-
tary concepts of time and space have begun to vacillate. Space is killed by the railways, and we
are left with time alone. … Now you can travel to Orleans in four and a half hours, and it takes
no longer to get to Rouen. Just imagine what will happen when the lines to Belgium and
Germany are completed and connected up with their railways! I feel as if the mountains and
forests of all countries were advancing on Paris. Even now, I can smell the German linden trees;
the North Sea’s breakers are rolling against my door.
7. Like most major technological systems, railroads fundamentally changed economic

and power structures. In the United States, for example, railroads validated the con-
tinental scale of the nation and restructured the economy from local business net-
works to national trusts (scale economies of national markets). With the railroad,
economic power passed to industrial firms from agriculture; more subtly, so did cul-
tural authority (Nye 1994).
8. The railroads fundamentally and radically shifted teleological dominance in Ameri-

can culture from Jeffersonian agrarianism, an Edenic teleology, to a technology-driven
New Jerusalem (Marx 1964; Nye 1994). This cultural schism replays itself even today,
with the sustainability and environmental discourses leaning toward the Edenic, and
the industrial, commercial, and science and technology communities supporting the
New Jerusalem teleology (Allenby, Allen, and Davidson 2007). Indeed, this split con-
tinues to appear in dialogs between the United States and the European Union; the
former sees technology as sublime, while for the latter, “nature” remains sublime
(Marx 1964; Grubler 1998). This cultural difference, although seldom recognized, goes
some way toward explaining such disagreements as those concerning agricultural
policy and genetically modified organisms, which the Americans see as an industrial
issue and the Europeans view as much more of an agrarian landscape issue.
This last point, the shift from technology as a challenge to agrarian Eden to technol-
ogy as a means to achieve the New Jerusalem (the “shining City upon a Hill”), is a
critical step in the relationship between technology and theology, but also in the
embrace of technology in the New World. Consider some selected sections from Walt
Whitman’s 1870 poem “Passage to India”:
Singing my days,
Singing the great achievements of the present,
Singing the strong light works of engineers,
Our modern wonders, (the antique ponderous Seven outvied,)
In the Old World the east the Suez Canal,
The New by its mighty railroad spann’d,
...............................
I see over my own continent the Pacific railroad surmounting every barrier,
I see continual trains of cars winding along the Platte carrying freight and passengers,
I hear the locomotives rushing and roaring, and the shrill steam-whistle,
I hear the echoes reverberate through the grandest scenery in the world,
...............................
After the seas are all cross’d, (as they seem already cross’d,)
After the great captains and engineers have accomplish’d their work,
After the noble inventors, after the scientists, the chemist, the geologist, ethnologist,
Finally shall come the poet worthy that name,
The true son of God shall come, singing his songs.
Then not your deeds only O voyagers, O scientists and inventors, shall be justified,
...............................
The whole earth, this cold, impassive, voiceless earth, shall be completely Justified,
...............................
Nature and Man shall be disjoin’d and diffused no more,
The true son of God shall absolutely fuse them.
136 Brad Allenby
“The true son of God shall absolutely fuse them”—in other words, the unity between
God, human, and nature in the form of New Jerusalem, the creation of the Shining
City on the Hill: this idea comes to the New World not from some ethereal realm, but
as the culmination of American technological prowess, evident in the railroad, among
other things. This is not technology as economic value, or as guarantor of national
security, but technology as salvation.
The railroad example makes several general principles of technological evolution
crystal clear. First, no technology system of any importance exists in a vacuum; rather,
there is a frontier, a salient, of coupled technology systems. Moreover, technology
systems are also coupled to many other human and natural systems; thus, a technol-
ogy of any significance will destabilize existing institutions and power relationships
as well as, to some degree, cultural assumptions. As a result, the technology will be
opposed by many. Accordingly, projecting the effects of technology systems before
they are actually adopted is not just hard but, given the complexity of the systems,
probably impossible. Thus, for example, the time structure that we moderns take for
granted was not the time structure of pre-railroad American agrarian society, nor was
it readily foreseeable before the railroad; it is a product of our technology. This
example raises a more subtle but equally important point: we are able to perceive our
world, and create our cultural constructs, only through the lens that our technology
provides. In a nontrivial sense, to be human has always entailed being an integration
of the biological and technological, and our minds function in part by enabling—
indeed, embracing—technologies as part of their cognitive processes (Clark 2003;
Allenby 2007b).
If the history of technological evolution is a warning, it is an inadequate one for
the wave bearing down on us. As suggested by the example of the railroads, techno-
logical change is always potent, but now we have not just one or two enabling tech-
nologies undergoing rapid evolution, but five: nanotechnology, biotechnology,
robotics, information and communication technology (ICT), and applied cognitive
science (together, NBRIC) (Garreau 2004). These technologies in some ways are the
logical end of the chapter in human history that began with the Greeks. Nanotechnol-
ogy extends human will and design to the atomic level, and represents the effective
culmination of humanity’s efforts to manipulate the material world. As for biotechnol-
ogy, McNeill (2000, 193–194) notes that “by the twentieth century, our numbers, our
high-energy technologies, and our refined division of labor with its exchange economy
made us capable of total transformation of any and all ecosystems. … In the twentieth
century we became what most cultures long imagined us to be: lords of the biosphere.”
ICT gives us the ability to create virtual worlds at will, and facilitates a migration of
functionality to information rather than physical structures. Thus, whereas money

used to be coins and paper bills, themselves mere symbols of value, now even that
physical premise is gone. Money is electrons somewhere in cyberspace, and financial
instruments have become so mathematical that no one can figure out anymore which
shell the risk is hidden under. Meaning and perception are always mediated by the
technologies available to a culture, but as the amount of information (and cultural
niches) proliferate in a world where information density increases daily, individuals
and virtual communities become increasingly isolated and realities increasingly con-
tingent and destabilized by noise level (Lyotard 1984).
Consider for a brief moment some of the implications of the NBRIC wave in just
one area, human biology and cognition, as an example of some of the possibilities.
At one extreme, some predict the achievement of “functional human immortality”
within fifty years, either as a result of continuing advances in biotechnology or as ICT
and computational power enable downloading of human consciousness into informa-
tion networks (Kurzweil 2005). While such predictions are viewed by most experts as
unlikely, there is a growing consensus that substantial extensions of average life spans,
with a high quality of life, are achievable in the next few decades (de Grey and Rae
2007). For example, the IEEE Spectrum, a mainstream technical journal, ran a series of
articles in 2004 on engineering and aging which concluded that using “engineered
negligible senescence” to control aging will allow average ages of well over 100 within
a few decades. It is interesting that this approach derives from applying principles of
electrical engineering to biological systems, thus reflecting the potency of integration
across previously well-defined disciplinary boundaries. From a governance perspective,
it is also noteworthy that, even though the scientists and technologists engaged
in this area perceive age extension as increasingly probable, those in other areas of
science—such as policy or the environmental and sustainability communities—remain
unaware of these possibilities, despite their obviously challenging implications (for
pension and old-age systems, material and energy consumption patterns, and the
dynamics of human/natural/built systems, for example). Thus, NBRIC turns not only
the earth but the human into a (highly reflexive) human design project, and substan-
tial changes in what it means to be human are probably inevitable (although specifics
are unpredictable).
N. Katherine Hayles, for example, in her aptly named book How We Became
Posthuman (1999), traces the evolution of the posthuman through the concepts of
homeostasis, then reflexivity, and finally virtuality. While Hayles is cautious about
the implications of this ongoing and accelerating process, some foresee enormous
potential. In an NSF report entitled Converging Technologies for Improving Human
138 Brad Allenby
Performance, Roco and Bainbridge (2003, ix) conclude: “With proper attention to
ethical issues and societal needs, converging technologies could achieve a tremendous
improvement in human abilities, societal outcomes, the nation’s productivity, and
the quality of life.” They continue:
Examples of payoffs may include improving work efficiency and learning, enhancing individual
sensory and cognitive capabilities, revolutionary changes in healthcare, improving both indi-
vidual and group creativity, highly effective communication techniques including brain-to-brain
interaction, perfecting human-machine interfaces including neuromorphic engineering, sustain-
able and “intelligent” environments including neuro-ergonomics, enhancing human capabilities
for defense purposes, reaching sustainable development using NBIC tools, and ameliorating
the physical and cognitive decline that is common to the aging mind. (Roco and Bainbridge
2003, ix)
Effects of technological convergence on the human are only one small area of research
and speculation; similar suites of possible scenarios are being developed in many other
areas. It is obviously premature to regard most of these predictions as anything more
than possible outcomes, useful scenarios and thought experiments to encourage devel-
opment of robust options in the face of uncertainty. Indeed, except for the easy cases
where particular applications of these core technologies are already in the process of
being commercialized, it is very difficult to determine how probable even the most
outré scenarios might be. But several relevant characteristics of technology systems
for trading zone theory are fairly clear.
Most importantly, technological change is not an isolated event, but represents
movements toward new, locally stable, earth systems states. These states integrate
natural, environmental, cultural, theological, institutional, financial, managerial,
technological, built, and human dimensions, and even construct our sense of time.
Technologies do not define these integrated earth systems states, except by conve-
nience, but technological evolution destabilizes existing sociotechnical clusters and
creates conditions leading to the evolution of new ones.
Moreover, the cultural effects of technology tend to become more disruptive the
more rapidly technology systems evolve, and the more fundamental the technology
involved. In this regard, current patterns of technological evolution are unprece-
dented. Previous technology clusters revolved around one or perhaps two evolving
technologies—say, rails and steam, or automobiles and petroleum. As touched upon
earlier, the constellation of nanotechnology, biotechnology, robotics, ICT, and cogni-
tive science marks a culmination of traditional technological evolution, for among
other things it reconstructs the earth, biology, and, indeed, even the human itself as
design spaces and, in doing so, renders contingent virtually all of what we have taken
to be fixed. Culturally, this stretches the bimodal distribution between a growing
global elite who, primarily through education and culture, are able to prosper under
such conditions, and those who are left behind. The latter have a strong tendency to
seek stability in outmoded ideologies and fundamentalist movements (whether radical
Islam, creationist Christianity, or fundamentalist environmentalism). These move-
ments are desperate responses to a world that such individuals see as having become
irrational, and, as the intense pace of change destabilizes patterns of belief and behav-
ior they invest with meaning, they see the world as profoundly challenging and fre-
quently evil.
Similarly, the American celebration of the technological sublime, whether expressed
in railroads or nuclear weapons, and whether celebrated out of fear or with joy, is an
example of technology used to express the human will to power. This is not just an
academic observation. Cultures that develop technology and, importantly, create
frameworks within which the technology can react upon itself, and so accelerate its
own evolution, thereby gain power over competitors. Because technologies create
such powerful comparative advantages between cultures, those cultures that attempt
to block technology usually will, all else being equal, eventually be dominated by
those that embrace it. This should not be interpreted as technological determinism,
for history clearly shows the contingency that affects human affairs (and thus the
systems coupled to them, which at this point are fairly inclusive). A Milošević, a Hitler,
a decision to drop the atomic bomb; more speculatively, a terrorist nuclear attack on
an American city—all these can significantly changed regional frameworks, at the
least. But, in general, the army with iron weapons dominates that with bronze; the
army with horses and stirrups dominates the foot soldiers; the gun dominates the
spear. Thus, it is likely that technological evolution will be difficult, if not impossible,
to stop, as some argue, for in a global economy with a number of nations jostling
for power, it is well understood that technology creates cultural comparative advan-
tage. The European Union can forswear genetically modified agricultural organisms
(GMOs), but they are still one of the most rapidly adopted technologies in agricultural
history; the Bush administration could refuse to support stem cell research, but it did
not stop Singapore and China, much less California, from doing so. Whether and
how emerging technologies, especially those that confer substantial economic or
security advantage, can be moderated in the age of global elites becomes an important
research question.
Implications for Trading Zone Theory
The previous discussion, attenuated though it is, makes it clear that society today faces
at least five different kinds of dramatically increasing complexity: (a) static complexity
140 Brad Allenby
(for example, increasing numbers of components, stakeholders, interactions among

different infrastructure, and linkages among them); (b) dynamic complexity, as these
factors interact in new and unanticipated ways, especially given the fundamentally
changing nature of ICT systems; (c) “wicked” complexity, arising from the need to
engineer and manage integrated human/natural/built systems that increasingly display
the reflexivity and intentionality of human systems and institutions); (d) scale, as we
realize we must begin to design, engineer, and manage integrated human/natural/built
earth systems at not just national but regional and global scales; and (e) ontological
complexity, as the global cultural framework of the Enlightenment fragments, is rene-
gotiated, and gives rise to information densities that defy the validity of single onto-
logical perspectives. This complexity has already had profound institutional
implications in our era: Marxist centralized planning in the Soviet Union and China
collapsed not from external conquest or even from Reagan’s vaunted spending race,
but rather because such explicit planning methods simply became incapable of man-
aging the complexity inherent in a modern industrial economy.
The ontological relativism introduced by accelerating complexity is particularly
important, for it reframes the assumptions of privilege (especially technocratic and
economic privilege) that underlie trading zone frameworks in most situations. In this
regard, an important element of this complexity is that it confirms an unavoidable
relationship between observer, frame of reference, and derivation of partial and con-
tingent truth from underlying complex systems. Consider a simple example. If I am
interested in the rates of crime in Phoenix, I am also implicitly defining the urban
system by its political boundaries. If, on the other hand, I am interested in water and
Phoenix, I am implicitly defining the system as including the Colorado River basin,
not to mention American water law, patterns of tourism that make golf courses
popular in Scottsdale, and xeriscaping initiatives. Yet in both cases the relevant marker
is “Phoenix.” What is happening, simply, is that my query to the system calls forth
from an underlying complex noumenal world a particular network that is responsive
to my query (in Kantian terms, the query acts to define an appropriate phenomenal
structure from a complex pattern of “things in themselves” that is not directly acces-
sible). Thus, complex urban systems can be thought of as interconnected, evolving
networks of networks, covering not just the familiar subjects of urban engineering—
built environments and infrastructure—but less physical ones such as technology
states, lifestyles, cultural constructs, economic evolution, and the like. For example,
while it is true that Chicago is a collection of buildings, roads, stores, and so forth, it
is also true that Chicago is the mechanism by which much of the American Midwest
was commodified.
The networks that are of interest in a particular situation will generally be deter-
mined not only by the evaluation of the noumenal reality, nor only by the particular
questions being asked about it, but through the interaction of the observer with the
noumenal system. There is a similarity to quantum mechanics here: what you perceive
when you look at the system is determined by the purpose for which you are observ-
ing it. The system itself always remains more complex than you are able to capture at
any one time. And the important corollary is that a complex system can be defined
only in terms of the reasons for which a definition is desired. The query identifies the
particular networks of the system that are relevant, and they in turn define the bound-
aries of the system for the purpose of the inquiry. This reflexivity complicates any
discussion of a complex system, of course, and reduces the value of standardized or
ideological approaches. Even in the cases where these integrated systems are com-
pletely built by humans—say, the Internet or urban systems or, eventually, designed
humans—their dynamics and evolutionary paths are neither planned nor determined
by humans, and their effects ripple broadly across many human, natural, and built
systems at many scales. These systems may be completely anthropogenic, but are not
understandable or transparent. When we design the human, to paraphrase Marx,
humans will make themselves, but they will not make themselves just as they please,
for our understanding and the complex nature of reality are not congruent, but
coupled weakly through our queries about reality.
If the first critical theme of the anthropogenic earth is complexity, then the second
is radical contingency, which coevolves with complexity. The accelerating evolution
of technology systems, especially ICT, combined with the postmodern fragmenting
of time, space, and culture, dramatically decreases the stability of all cultural constructs
(Allenby 2006; Allenby, Allen, and Davidson 2007). In particular, the contingency of
the anthropogenic earth is radical because, in addition to making the social and cul-
tural landscapes more unstable, it also renders that which looks out—the self and
human individuality—more contingent. The dramatic increase of fundamentalism
across most belief systems and in most societies reflects, in part, an effort to create a
stable ground; it is an effort that will fail, at least for the elite for whom transhuman-
ism is already a reality. Marx’s prediction—or curse, depending on your viewpoint—
comes true: all that is solid melts into air. Note that this does not mean that the
postmodern solutions of absolute solipsism and relativism are valid (Lyotard 1984); it
simply means that if our mental models and cultural constructs are to be adaptive,
they must embrace, and manage, their own contingency.
Moreover, given the evolution in networked cognitive systems that currently char-
acterize the “digital native” generation, it is likely that this contingency will be heavily
142 Brad Allenby
influenced by technology (Prensky 2001). The dialectic of technological evolution is

often taken to require a conflict between technology and the human, as if each were
a separate domain. The Enlightenment romantics had their Frankenstein model, and
it remains powerful today (as in Greenpeace’s Frankenfood PR campaign, or the movie
The Matrix). If history is any guide, the conflict between technology and the human
is, at most, a temporary opposition: it is not that the human and the technological
will clash, and one will emerge victorious; rather, what is already happening is that
the two are merging. This does not mean profound changes won’t occur, especially
in older concepts of what constitutes “the human.” Nor does it mean that we won’t
see varietals of humans—as, indeed, the “digital natives” who are comfortably embed-
ded in their ICT networks may already be. At the level of “nature,” it means that we
should expect integrated human/natural/built earth systems, rather than those we
currently romanticize. Indeed, some current Internet mash-ups, where representations
of the real world are mixed online with virtual representations of data sets or imagi-
nary spaces, are already going in that direction.
This radical contingency, extending beyond external systems to core belief struc-
tures and cultural concepts defining the human, makes ideological frameworks—the
definitional basis of many communities potentially involved in trading zones—even
more problematic than usual. Many groups, from deep greens (ideological environ-
mentalists) to Marxists, from religious conservatives opposing modernity to scientists
using climate change evidence to make moral arguments, cling to worldviews and
ideologies embedding increasingly anachronistic teleologies and assumptions about
totalizing values. There are obvious reasons for doing so: such posturing engages their
base; provides emotional force for otherwise unpopular policies; enables social engi-
neering by committed minorities; turns complex questions of fact into simplistic
black-and-white scenarios; and, in many cases, reflects and validates the rejection of
modernity. But ideological approaches of all kinds are particularly problematic at the
dawn of the anthropogenic world, characterized as it is by exceedingly rapid and
profound change in fundamental relationships and systems, and involving integrated
human/natural/built systems of extraordinary complexity. In such a context, there are
four aspects of ideology that render it especially dysfunctional. First, any ideology is
necessarily a simplification of reality; in fact, that’s usually an important part of its
mass appeal. Second, the elements and structure of this simplification necessarily lie
in the past, not the future, and thus embed assumptions and implications that are
necessarily increasingly anachronistic in a period of rapid and discontinuous change.
Third, ideology creates an “ends justify the means” mentality; almost by definition,
the power of the Idea trumps the messy and contingent real world. Thus, it is char-
acteristic of many ideologies that they posit a vision of utopia, the achievement of
which is worth the sacrifices, usually imposed by the ideological group on others.
Finally, as part of the elevation of the Idea over the real, ideology also cuts off infor-
mation transfer and dialog, and is profoundly antidemocratic, anti-intellectual, and
antirational (although, ironically, ideologies are creatures of the intelligentsia). It is
not, then, just that ideologies are generally bad, although many of them seem to be
in application, as any familiarity with the twentieth century would confirm (Conquest
2000); rather, it is that ideologies are especially bad in a period of rapid, discontinuous,
and fundamental change at a global, multicultural scale.
Taken together, these factors support an important role for trading zones in facili-
tating dialog about many classes of complex and difficult issues. But they also indicate
some important limits to the concept. First, not everyone wants to communicate;
indeed, ignorance is a necessary comfort for many, and is usually benefiting someone.
Maintaining ignorance may be Machiavellian, or it may simply reflect the fact that,
to those in thrall to particular ideologies, open communication is not just unnecessary,
since they believe the answers are already known, but is actually evil. (I’m reminded
of the motto I saw outside a New England fundamentalist church several years ago:
“An open mind is the devil’s playground.”) If change, complexity, and contingency
drive one to fundamentalism, it is highly unlikely that dialog will resolve any of the
underlying formative pressures.
Second, the idea of trading zones is limited in that it is not transcendent, but is
itself a reflection of general Enlightenment and Western values supporting participa-
tory processes and liberal democracy which are not necessarily universally shared. That
they are important values in a multicultural world is a valid assertion; that they
embody a certain worldview and associated cultural constructs cannot be ignored.
In particular, it may be that current conditions undermine the viability of liberal
democratic processes as a whole. The increasingly potent and global fundamentalist
backlash that is occurring in virtually all major religions, as well as those belief
systems—environmentalism, sustainability—that for many people, especially in
secular societies, now begin to serve theological purposes, is not just random opposi-
tion to modernity. Rather, it reflects the fact that, as rates of technological change
accelerate, increasing numbers of people in every society are effectively disenfran-
chised. They are incapable of keeping pace with continuing change, unable to inte-
grate into the information webs that increasingly define human cognition, and aghast
at the changes in lifestyle, income distribution, relative power relationships, and
changes in sexual and family roles and structures that have resulted. And, importantly,
these groups have not yet understood the degree to which their fundamental values
144 Brad Allenby
are rendered contingent by that selfsame progress. Thus, the ever more rapid evolu-
tion of the anthropogenic earth, and accelerating technological change, can only
increase opposition to change. For those for whom Enlightenment representative
democracy is an important value, then, this change creates a difficult conundrum,
for the more it succeeds, the more it creates an activist opposition that hobbles
change in democratic cultures, giving the advantage to cultures where the elite, who
benefit from technological evolution generally, are able to exercise control. Impeding
continuing technological evolution in democratic societies, does not, after all, mean
that technological evolution doesn’t happen; merely that it happens in societies that
are not thus restricted. Over time this would reduce the technological competency,
and therefore the comparative advantages, of liberal democratic societies, while
authoritarian societies become increasingly competent because of greater willingness
to support technological change regardless of the cost. Thus, it may ironically be the
case that the changes generated by the Enlightenment and the rise of democratic
and secular societies will lead to rates of change, complexity, and radical contingency
that result in the end of the great Enlightenment project of liberal democratic socie-
ties. Should such a scenario develop, it would obviously undermine governance
mechanisms, such as trading zones, predicated on greater communication, transpar-
ency, and openness.
The third limitation of the concept of trading zones is that the existence or creation
of trading zones does not in itself overcome obsolete mental models and worldviews.
Thus, to the extent that increasingly dysfunctional institutions reflect a lack of vision
and understanding, and inability to grapple with complexity and radical contingency,
trading zones not only do not provide substantive relief but, by appearing to be a
validating process, may in fact impede intellectual and cultural adaptation. Trading
zones enable construction of adaptive knowledge when that knowledge is at least
implicitly already in the system; when all communities involved in the trading zone
suffer from increasingly anachronistic ideologies and mental models, trading zones
do not clearly help to create viable institutional responses. This is especially true
because trading zones are dialogs among human communities; if, as suggested in this
chapter, the challenge is to learn to dialog among natural, built, and human systems—
and it is also recognized that the human systems are increasingly contingent over
time—a much broader concept of dialog is engaged. From the perspective of develop-
ing expertise in earth systems engineering and management (ESEM), moreover, it is
not immediately apparent that the challenge of creating appropriate analytical frame-
works by dialoging with an ever more complex noumenal reality is either facilitated
or necessarily encouraged by trading zone methodologies.
It is, however, somewhat unfair to challenge trading zone theory by asking it to

overcome the cumulative burdens of cultural obsolescence. Indeed, even for many
complex earth systems where conflicting ontologies are in play, trading zone methods
may be quite desirable. One example already discussed is the Everglades, where the
adaptive management process to some extent already resembles a trading zone; one
can also posit that the high modernist disaster in the Aral Sea was in part a result of
a lack of sufficient dialog. Indeed, it would seem that so long as the systems at issue
are culturally bounded and do not engage foundational contingency, trading zone
methods, appropriately applied, could be quite useful. But in other cases, such as the
growing debate over transhumanism (Allenby 2007b), trading zone methods may be
less appropriate, because the challenge raised by the system undermines the very idea
of stable ontological and teleological frameworks that enable trading zones to exist.
Where everything is contingent, dialog based on existing mental models and cultural
constructs may not be able to reach the depth required for true understanding. Accord-
ingly, trading zones may need to be augmented by dramatic changes in how the
ethical responsibility and rationality of the individual are framed, and the further
development of the new integrative field of ESEM. In this view, trading zones are an
important response to a difficult principle of ESEM—the need to facilitate transparent
and democratic dialog—but they are not, in themselves, adequate for a full under-
standing of ESEM or the anthropogenic earth. Accordingly, this discussion closes by
addressing these two complementary responses to the challenges of complexity, con-
tingency, and accelerating change.
The Personal Dimension: Toward a New Authenticity
The combined challenges of unprecedented change across natural, human, and built
systems; complexity; and radical contingency are secondarily a challenge to gover-
nance mechanisms, such as trading zones. But, first and foremost, they constitute a
challenge to individual humans and their ability to cope and evolve to meet new
conditions. For individual humans must, at the very least, develop appropriate under-
standing and frameworks if relevant and sophisticated trading zones are to be devel-
oped and appropriately applied. Thus, this discussion will close by suggesting at least
some characteristics associated with the authenticity that the rise of the anthropogenic
earth, with its accelerating and constant change, increasing complexity, and radical
contingency, calls forth (while the full meaning of “authenticity” will become clearer
over time and through discussion and dialectic, the concept may be understood as
including an acceptance of the complexity, radical contingency, and unpredictability
146 Brad Allenby
of the anthropogenic earth, and the ad hoc, constructed, and always partial essence
of any explicit framework created to explain, or help design and manage, the complex
adaptive human/natural/built systems that characterize it).
1. Following the existentialist formulation (and, for that matter, going back to Socrates’
injunction to “know thyself”), an authenticity necessary for our times will require as
a first element a recognition and acceptance of the world as it is, not as various ideolo-
gies would wish it to be. Thus, simplistic ideologies and worldviews must be rejected.
2. The human condition must also be accepted, in that the anthropogenic earth
requires individuals to accept the validity of their condition and cognitive networks
for themselves, while simultaneously recognizing them as contingent and stochastic
in a world characterized by mutually exclusive but equally valid ontologies. Each
individual represents local truth and global relativism.
3. Authenticity also requires acceptance of the epistemological and existential impli-
cations of complex adaptive systems, in that any perceptual or cognitive network, or
understanding of a complex system, is created by the query posed to the system, and
thus embodies unavoidable reflexivity between the system and the cognitive network,
and implies the contingency and incompleteness of any particular perspective on a
complex adaptive system.
4. Given proposition 3, authenticity demands that we must have the integrity to
create appropriate queries, since they will structure the cognitive networks within
which we operate. Substituting wistful fantasies for honest query and thus construc-
tion of our local realities, or game-playing the query process to create ideologically
predetermined local realities, must be rejected as profoundly inauthentic.
5. Authenticity requires that we accept the condition that meaning, truth, and values
do not arise from first principles, but are functions of network state, and thus are
contingent and continually regenerated in a reflexive dialog between cognitive systems
posing queries to, and thus generating configurations of, external complex adaptive
systems.
6. Following propositions 4 and 5, authenticity requires accepting as the human
condition the challenge that you must distrust the most that which you most believe.
Meaning and truth arise from the dialectical process of their continued rejection.
7. Authenticity requires accepting rationality as partial and constructed, an interplay
between different and contingent ontologies and partial structures of underlying
complex adaptive systems, congealed intentionality and cognition, and institutional
and network dynamics. A similar stance must be taken toward institutions, or, indeed,
any cognitive network. In doing so, however, the mistake of slipping into a solipsistic
relativism must be avoided, for that goes too far and becomes its own form of inau-
thenticity, a mere mechanism to avoid responsibility.
8. Even though the macroethics of complex adaptive systems are beyond the level of
the individual, authenticity requires that each individual, operating in good faith,
participate in establishing institutional capabilities to dialog with such systems, be
they technological, environmental, biological, cultural, or social.
9. As a reflection of the increasing human role in, and responsibility for, integrated
human/natural/built earth systems, authenticity requires thoughtful rejection of ide-
ologies and frameworks characteristic of the first Enlightenment, and active move-
ment toward reinvention of the Enlightenment to suit a profoundly multicultural,
and much more complex, world. This rejection must be thoughtful, for out of the first
Enlightenment must be created a second that embodies the best elements of the first
while enabling responses to new conditions, but we must also reject those elements
which now constitute cultural or temporal imperialism, or are too simplistic for the
systems that characterize the Anthropocene.
10. Finally, authenticity requires understanding that the individual is a contingent
framework that has worked well in the past, but is increasingly partial in a complex
world characterized by cognitive networks extending across technological, biological,
and human systems, and the evolution of transhuman variants that is already well
under way. Thus, authenticity demands acceptance of cognition as increasingly
involving production of emergent systems characteristics at levels higher than the
individual.
11. This authenticity does not reject theology, but redistributes domains between the
theological and the human in ways that may be very difficult for many individuals
to accept. The strength to accept such shifts, while at the same time not succumbing
to mere relativism, is an important element of the authenticity required.
With knowledge of the anthropogenic earth comes an existential crisis, as the

honest perception demanded by authenticity reveals a chaotic, unpredictable, highly
problematic planet in the throes of anthropogenic change, with a complexity that
cannot be adequately addressed by either existing intellectual tools or language itself.
Each individual is profoundly ignorant, and strives hard to remain ignorant even of
his or her ignorance; naiveté and willful perceptual and intellectual blindness become
comfortable characteristics of discourse. And the result is a fleeing into ideology,
random myths, and stories, the creation of mental models that simplify reality into
manageable fantasy, and reduce perception until it no longer threatens. This is under-
standable, but it is profoundly inauthentic—a flight from freedom, from responsibility,
148 Brad Allenby
from integrity. As Sartre said in the context of the individual, “Man is condemned to
be free.” And this is a far more daunting challenge in the context of an anthropogenic
world that we, having created it, now want to pretend not to see. For now this
freedom, from whence rises moral obligation, is neither comfortable nor, sometimes,
even bearable. But it is the freedom demanded by the historical moment, and we
cannot delegate it.
He, only, merits freedom and existence
Who wins them every day anew.
Goethe, Faust, 11575–11576
The Institutional Perspective: Earth Systems Engineering and Management
The continued stability of the complex, information-dense, highly integrated human/

natural/built systems that characterize the anthropogenic earth requires the develop-
ment of the ability to rationally design, engineer, and construct, maintain and manage,
and reconstruct systems in a highly integrated and profoundly multicultural fashion—
in short, an ESEM capability (Allenby 2000; Allenby 2007a). ESEM is a field of study:
How does one learn the appropriate skills to understand these systems, and to inter-
vene in and dialog with them in such a way as to evolve appropriate paths? Very
much like some systems engineering, it is also a posture: How does the ESEM practi-
tioner stand outside both herself and the system to imagine and help implement viable
policies and practices, while at the same time realizing that one is simultaneously a
part of that system, and responding to the reflexivity and instability that accordingly
results? ESEM is also a framework within which the complex and rapidly evolving
systems of the Anthropocene can be addressed responsibly, ethically, and rationally.
Conversely, ESEM is not the application of routine engineering and management
practices to complex systems, as happened with the Aral Sea; such an approach is in
fact an important failure mode that must be understood and avoided.
While developing its field and practice constitute a significant challenge, ESEM can
draw on experience from a number of nascent areas of study and practice, in addition
to the growing literature on trading zones. From a technical perspective, these would
include industrial ecology methodologies such as life cycle assessment, design for
environment, and materials flow analysis (Ayres and Ayres 2002; Graedel and Allenby
2003), as well as experience gained in systems engineering (Pool 1997; Hughes 1998).
From a managerial perspective, it draws on the literature about managing complexity
and “learning organizations” (Senge 1990; Gharajedaghi 1999). From a resource man-
agement perspective, it draws on the adaptive management literature developed
primarily by scientists and ecologists working on complicated resource regimes and

regional hydrological phenomenon (Gunderson, Holling, and Light 1995; Berkes and
Folke 1998). It also draws on areas that may be less familiar to technocratic profes-
sionals, such as literature on planning, on the sociology of technology, and on social
construction (Berger and Luckmann 1967; Bijker, Hughes, and Pinch 1997).
Based on these discourses, one can develop a tentative and preliminary, albeit
instructive, set of initial ESEM principles, including the following:
1. Given our current level of ignorance, one should only intervene when necessary, and
then only to the extent required, in complex systems. This follows from the obvious need
to treat complex adaptive systems with respect, since one can seldom predict their
future paths and reactions to inputs; moreover, under such circumstances, minimal
interventions reduce the possibility and to some extent the potential scale of unfore-
seen and undesirable system responses. Note this is very different from the so-called
Precautionary Principle, which essentially states that if one cannot know the risks
posed by a new technology or practice, one should not introduce it. Given that one
can never know future paths in complex adaptive systems marked by high degrees of
reflexivity, it follows that one also cannot know the risks or benefits posed by any
new technology. Thus, either a technology is trivial, in which case precautions are
unnecessary, or it is unpredictable, in which case the strict application of this formula-
tion simply becomes a rationale for rejecting technological evolution.
2. Major shifts in technological systems should be evaluated before, rather than after, imple-
mentation of policies and initiatives designed to encourage them. This principle sounds
straightforward, but activists, the public, and policymakers frequently fall in love with
particular technologies, making study of their potential downsides both economically
and, frequently, socially infeasible (for funding intended to question what “everyone
knows is right” is seldom available). Consider, for example, the current infatuation
with biofuels, particularly the massive effort by the United States to produce corn-
based ethanol at large scale as a major component of its energy economy, despite
serious questions about how its production at a large scale might cause changes in
land use and water management, further disruption of the nitrogen and phosphorus
cycles, and major increases in food prices. An example from another field is e-
commerce in its various guises, such as B2B (business-to-business) or B2C (business-
to-consumer), which continues to expand rapidly despite a profound lack of
understanding of its potential social, environmental, and economic implications
(Allenby 2000). The time to begin studying these issues is now, while the technologies
are being developed, rather than later, when the effects may well have already occurred,
and changes to technology become much more difficult due to lock-in, development
150 Brad Allenby
of standards, network economics, and other phenomenon (Castells 2000). The point
is not, of course, that technology shifts may not be beneficial; the point is to improve
their design and management within the real-world frameworks within which they
evolve.
3. As demonstrated above, one characteristic of complex systems is that the network
that is relevant to a particular analysis is called forth by that analysis. Accordingly, it
is critical to be aware of the particular boundaries within which one is working, and to be
alert to the possibility of logical failure when one’s analysis goes beyond the boundaries. For
example, to do a study of New York City’s water supply by considering only the five
constituent boroughs of the city would result in a flawed assessment, because the
system being analyzed (water provision to the city) is not adequately mapped by the
political boundaries of the city. The relevant network and the boundary selected are
discordant.
4. It is critical in the anthropogenic world to understand that the actors and designers
are also part of the system they are purporting to design, creating a reflexivity that makes the
system highly unpredictable and, to some extent, perhaps unstable as well. Thus, it is impor-
tant that activities at the levels of the emergent behaviors of these complex systems
must be understood as processes and dialogs, rather than simply as problems to be
solved and forgotten. This bifurcates engineering: most engineering still involves
artifacts, but earth systems engineering and management requires ongoing and highly
sophisticated dialogs with the systems at issue. For example, a civil engineer may still
design a levee in the Everglades and walk away, but an ESEM practitioner responsible
for the Everglades itself must maintain a constant dialog with that system (the levee,
for example, may have unforeseen effects on water flow in other parts of the Ever-
glades, and thus have to be modified or enhanced over time).
5. The conditions characterizing the anthropogenic earth require democratic, transparent, and
accountable governance, and pluralistic decision-making processes. Virtually all ESEM initia-
tives raise important scientific, technical, economic, political, ethical, theological, and
cultural issues in the context of an increasingly complex global polity. Given the need
for consensus and long-term commitment, the only workable governance model is
one that is democratic, transparent, and accountable (Habermas 1975; Rorty 1989). It
is here that the value of trading zone theory to the practice of ESEM becomes
apparent.
6. It follows from the complexity of the systems involved, our relative ignorance, and
the recognition of engineering as process that continual learning at both the personal
and institutional level must be built into project and program management. Some experience
with this approach already exists. In the case of technological systems and their man-
agement, for example, existing “high-reliability organizations” such as aircraft carrier
operations or well-run nuclear power plants usually have explicit learning structures
(Pool 1997). In the case of complex natural resource management challenges, such as
the Aral Sea, the Everglades, the North American Great Lakes, or particular forest or
fishery resources, the adaptive management approach is heavily dependent on con-
tinual learning (and the relevant literature identifies cases where lack of this ability
makes policy and social institutions quite brittle as change occurs) (Gunderson,
Holling, and Light 1995; Berkes and Folke 1998). Indeed, the study of mixed human/
natural/built systems—human individuals and technologies acting as integrated cog-
nitive networks in performing complex tasks such as navigation (Hutchins 1995)—sug-
gests that this learning task is not just difficult at a pragmatic and funding level, but
requires rethinking our educational process itself from a very basic level.
7. When working with unpredictable systems, it is important to track both desired
and, particularly, undesired outcomes, in part by establishing metrics that determine
whether the system is indeed moving along an appropriate path to achieve the desired out-
comes. While it may be that not all potential behaviors can be identified ab initio, such
metrics are part of the dialog with the system, and, especially, the development of a
capability to anticipate undesirable behaviors before they become locked in.
8. Along these lines, premature lock-in of system components should be avoided where
possible. In complex systems, artifacts, technologies, and choices among alternatives
can quickly get locked in, depending on the structure of the particular system or
network. Moreover, tightly coupled networks are more resistant to change than loosely
coupled networks. These characteristics of complex systems explain, for example, why
changes to regulations of pollution control equipment are more easily accomplished
than changes to regulations of product design or manufacturing processes. In the
former case, the technology is only loosely coupled to underlying product and
manufacturing networks, and thus can be changed with only minimal implications
for other aspects of the product and manufacturing networks; lock-in exists but is not
significant. In the latter case, however, product design and manufacturing processes
are complex, tightly coupled, interdependent networks, which means that changes
and shifts in state propagate throughout the system. Lock-in of changes to such
systems occurs more easily, and is more irreversible, in the sense that once the inter-
ventions are implemented, it is much more difficult to change them.
9. Similarly, whenever possible, engineered changes should be incremental and reversible,
rather than fundamental and irreversible. Accordingly, lock-in should be avoided (because
it leads to irreversibility), in part by assuring that, where couplings to other networks
152 Brad Allenby
do exist, they are designed to be as loose as possible and as few as possible. After all,
if the system begins to behave in an unanticipated and undesired way because of the
action taken, easy reversibility is a very valuable characteristic. Moreover, as a funda-
mental technology system is implemented, scale-up should anticipate nonlinear
behavior on the part of the system; emergent characteristics should be assumed to be
the rule, not the exception.
10. Earth systems engineering and management projects should aim for resiliency, not
just redundancy, in design. Redundancy provides backup capability in the event that a
primary system fails, and is commonly designed into high reliability systems such as
jet airplanes. Redundancy assumes, however, that the challenge to the system is of a
known variety. Resiliency, on the contrary, is the ability of a system to resist degrada-
tion and, when it must degrade, to do so gracefully even under unanticipated condi-
tions. Neither redundancy nor resiliency is easy to achieve. Designing redundancy
requires enough knowledge of the system to identify likely failure modes. Designing
resiliency also requires a high degree of knowledge, in part because resiliency is not a
global characteristic of a complex system, but is generally defined in terms of capabil-
ity to resist a particular class of insults or perturbations (Allenby and Fink 2005). Thus,
it is important when working with a system—say, trying to help an urban region
improve its ability to resist and respond to terrorism, while at the same time creating
better quality of life for its inhabitants—to define resiliency in appropriate terms, so
that policies can be evaluated and resource allocations made with some idea of associ-
ated costs and benefits.
Conclusion
The demographic, economic, cultural, and technological changes occurring since the
Industrial Revolution have given rise to a new and challenging phenomenon: a planet
dominated by the activity and culture of a single species. This anthropogenic earth is
a challenge because it is characterized by integrated human/natural/built systems that
are highly complex and rapidly changing, and that thus create instability and radical
contingency across not just physical systems but also cultural and psychological
systems. Trading zone theory is a helpful contribution to the integrated response of
earth systems engineering and management. Although this field is still young and
difficult to define, in part because of the complexity of the systems with which it deals,
current knowledge is adequate to begin to develop realistic personal and institutional
ESEM frameworks. While these are not yet comprehensive, they are certainly grounded
and realistic enough to enable practices, policies, and behaviors that are more respon-
sive than current options.
Notes
1. Parts of this and the following sections are drawn from the author’s Templeton Fellow lecture,
“From Human to Transhuman: Technology and the Reconstruction of the World,” presented at
Arizona State University, Tempe, October 22, 2007.
References
Allenby, B. R. 2000. Earth Systems Engineering and Management. Technology in Society 19

(4):10–24.
Allenby, B. R. 2005. Reconstructing Earth. Washington, DC: Island Press.
Allenby, B. R. 2006. The Real Death of Environmentalism. Environmental Quality Management 16

(1):1–10.
Allenby, B. R. 2007a. Earth Systems Engineering and Management: A Manifesto. Environmental

Science and Technology 41 (23):7960–7965.
Allenby, B. R. 2007b. From Human to Transhuman: Technology and the Reconstruction of the
World. Templeton Fellow Lecture, Arizona State University, Tempe, October 22. http://www
.cspo.org/documents/Templeton%20Public%20Lecture%20-%20FROM%20HUMAN%20TO%20
TRANSHUMAN.pdf (accessed March 2010).
Allenby, B. R., D. Allen, and C. Davidson. 2007. Sustainable Engineering: From Myth to Mecha-
nism. Environmental Quality Management 17 (1):17–26.
Allenby, B. R., and J. Fink. 2005. Toward Inherently Secure and Resilient Societies. Science
309:1034–1036.
ARAL. 2003. Aral Sea Loss and Cotton (Case Number 185). http://www1.american.edu/ted/
aral.htm (accessed March 2010).
Ayres, R. U., and L. W. Ayres. 2002. A Handbook of Industrial Ecology. Cheltenham, UK: Edward
Elgar.
Berger, P. L., and T. Luckmann. 1967. The Social Construction of Reality: A Treatise in the Sociology
of Knowledge. New York: Anchor Books.
Berkes, F., and C. Folke, eds. 1998. Linking Social and Ecological Systems: Management Practices and
Social Mechanisms for Building Resilience. Cambridge: Cambridge University Press.
Bijker, W. E., T. P. Hughes, and T. Pinch, eds. 1997. The Social Construction of Technological Systems.
Cambridge: MIT Press.
154 Brad Allenby
Castells, M. 2000. The Rise of the Network Society. 2nd ed. Oxford: Blackwell Publishers.
Clark, A. 2003. Natural Born Cyborgs: Minds, Technologies, and the Future of Human Intelligence.
Oxford: Oxford University Press.
Collins, H., R. Evans, and M. Gorman. 2007. Trading Zones and Interactional Expertise. Studies
in History and Philosophy of Science 38:657–666.
Conquest, R. 2000. Reflections on a Ravaged Century. New York: W. W. Norton.
Cronon, W. 1991. Nature’s Metropolis: Chicago and the Great West. New York: W. W. Norton.
Cronon, W., ed. 1995. Uncommon Ground: Rethinking the Human Place in Nature. New York:
W. W. Norton.
de Grey, A., and M. Rae. 2007. Ending Aging: The Rejuvenation Breakthroughs That Could Reverse
Human Aging in Our Lifetime. New York: St. Martin’s Press.
FAO (Food and Agriculture Organization of the United Nations). 1998. Time to Save the Aral
Sea? The Destruction of This Central Asian Lake Is a Textbook Example of Unsustainable
Development. http://www.fao.org/ag/magazine/9809/spot2.htm (accessed March 2010).
Freeman, C., and F. Louca. 2001. As Time Goes By: From the Industrial Revolutions to the Information
Revolution. Oxford: Oxford University Press.
GAO (U. S. General Accounting Office). 2003 (March). South Florida Ecosystem Restoration: Task
Force Needs to Improve Science Coordination to Increase the Likelihood of Success. Report to Subcom-
mittee on Interior and Related Agencies, Committee on Appropriations, House of Representatives.
Washington, DC: GAO.
Garreau, J. 2004. Radical Evolution. New York: Doubleday.
Gharajedaghi, J. 1999. Systems Thinking: Managing Chaos and Complexity. Boston: Butterworth
Heinemann.
Graedel, T. E., and B. R. Allenby. 2003. Industrial Ecology. 2nd ed. Upper Saddle River, NJ:
Prentice-Hall.
Grubler, A. 1998. Technology and Global Change. Cambridge: Cambridge University Press.
Gunderson, L. H., C. S. Holling, and S. S. Light, eds. 1995. Barriers and Bridges to the Renewal of
Ecosystems and Institutions. New York: Columbia University Press.
Habermas, J. 1975. Legitimation Crisis. Trans. T. McCarthy. Boston: Beacon Press.
Harvey, D. 1996. Justice, Nature and the Geography of Difference. Cambridge, MA: Blackwell
Publishers.
Hayles, N. K. 1999. How We Became Posthuman. Chicago: University of Chicago Press.
Hughes, T. P. 1998. Rescuing Prometheus. New York: Pantheon.
Hutchins, E. 1995. Cognition in the Wild. Cambridge: MIT Press.

IEEE Spectrum. 2004. Engineering and Aging. IEEE Spectrum 41 (9):10, 31–35.
Kaiser, J. 2001. NRC Panel Pokes Holes in Everglades Scheme. Science 291:959–961.
Kloor, K. 2000. Everglades Restoration Plan Hits Rough Waters. Science 288:1166–1167.
Kurzweil, R. 2005. The Singularity Is Near. New York: Viking.
Lawler, A. 2005. Reviving Iraq’s Wetlands. Science 307:1186–1189.
Lyotard, J. 1984 [1979]. The Postmodern Condition: A Report on Knowledge. Trans. G. Bennington
and B. Massumi. Minneapolis: University of Minnesota Press.
Marx, L. 1964. The Machine in the Garden: Technology and the Pastoral Ideal in America. Oxford:
Oxford University Press.
McNeill, J. R. 2000. Something New under the Sun. New York: W. W. Norton.
NASA. 2008. Earth at Night. Antwrp.gsfc.nasa.gov/apod/ap020810.html (accessed February

2008).
Nature. 2003. Welcome to the Anthropocene [editorial]. Nature 424:709.
Needham, J. 1988 [1954]. Science and Civilization in China. Vol. 1. Cambridge, UK: Cambridge
University Press.
NRC (National Research Council). 2007. Progress toward Restoring the Everglades: The First Biennial
Review. Washington, DC: National Academy Press.
Nye, D. E. 1994. The American Technological Sublime. Cambridge, MA: MIT Press.
Pala, C. 2006. Once a Terminal Case, the North Aral Sea Shows New Signs of Life. Science 312:183.
Pool, R. 1997. Beyond Engineering: How Society Shapes Technology. Oxford: Oxford University Press.
Prensky, M. 2001. Digital Natives, Digital Immigrants. Horizon 9 (5):1–6.
Roco, M. C., and W. S. Bainbridge, eds. 2003. Converging Technologies for Improving Human Perfor-
mance. Dordrecht: Kluwer Academic Publishers.
Rorty, R. 1989. Contingency, Irony, and Solidarity. Cambridge: Cambridge University Press.
Rosenberg, N., and L. E. Birdzell, Jr. 1986. How the West Grew Rich: The Economic Transformation
of the Industrial World. New York: Basic Books.
Schivelbusch, W. 1977. The Railway Journey: The Industrialization of Time and Space in the Nine-
teenth Century. Berkeley: University of California Press.
Schmidt, K. F. 2001. A True-Blue Vision for the Danube. Science 294:1444–1447.
Schrope, M. 2001. Save Our Swamp. Nature 409:128–130.
Scott, J. C. 1998. Seeing Like a State: How Certain Schemes to Improve the Human Condition Have
156 Brad Allenby
Senge, P. M. 1990. The Fifth Discipline. New York: Doubleday.
Shapiro, J. 2001. Mao’s War against Nature: Politics and the Environment in Revolutionary China.
Cambridge: Cambridge University Press.
SUSF (State University System of Florida). 2003. Reclaiming the Everglades: Everglades Timeline.
http://everglades.fiu.edu/reclaim/timeline/index.htm (accessed January 7, 2010).
UNEP-IFAS (United Nations Environment Programme and International Fund for the Aral Sea).
2008. Aral Sea. http://www.unep.org/dewa/vitalwater/article115.html (accessed March 2010).
8 The Evolution of a Trading Zone: A Case Study of the Turtle
Excluder Device
Lekelia D. Jenkins
Introduction
Sea turtles are among the best-known marine endangered species, and shrimping is
the most profitable U.S. fishery. Thus, the incidental death of sea turtles in shrimp
trawls (a problem generally known as bycatch) resembled the clash of two juggernauts,
and became one of most controversial problems ever confronted by U.S. fisheries
management. The stakeholders were numerous and diverse, from politicians to school-
children, fishers to environmentalists, and scientists to fisheries managers. Each group
had differing opinions of how best to resolve the problem, and cooperation between
groups would be critical to a successful solution, which came in the form of a turtle
excluder device (TED)—a type of turtle escape hatch in the shrimp net.
The concepts of trading zones and interactional expertise are useful for character-
izing the types of engagements between stakeholder groups and how they contributed
to the final resolution of the sea turtle bycatch issue. In this paper, I will describe
examples of fractionated and enforced trading zones as they played out in the case
study. In addition, I will offer evidential support for the hypothetical portrayal of the
evolution of a trading zone proposed by Collins, Evans, and Gorman (2007) by orga-
nizing the case study within the trading zone model, thus showing the evolution from
institutional power to boundary object to interactional expertise. I will also show how
a trading zone can split and diverge into separate trading zones. For each step of the
evolution, I will describe the forces that drove these transitions. Finally, I will present
an adapted trading zone model that is conceptually a better fit for the TED case study.
Methods
I examined the TED case study between 1976—when research began with the goal of
reducing sea turtle bycatch—and 1998. The final year included in the TED case study
158 Lekelia D. Jenkins
was the last year of TED development before major changes in TED regulations.1 I
gathered data for this study by examining inventions, conducting interviews, and
analyzing documents. I conducted thirty-three on-site semistructured and unstruc-
tured personal interviews with key informants.2 These interviews mostly occurred
during three two-week-long trips in June 2003, August 2003, and September 2003. The
sample population consisted of representatives from stakeholder groups, including
federal and state policymakers and managers, scientists, inventors, as well as represen-
tatives of the fishing industry and environmental organizations. I initially established
a sample frame using a purposive sample of prominent individuals frequently men-
tioned in the literature pertaining to the study.3 The purposive sample led to a snow-
ball sample: informants were asked to name other individuals who were knowledgeable
about the case study, and many of these were then interviewed.4
I also collected hundreds of documents, including government reports, research
records, workshop reports, panel reports, memos, personal letters, and educational
videos and pamphlets from the key informants’ archives. I analyzed the text of the
interviews and documents in the spirit of grounded theory (Strauss and Corbin 1998),
allowing theories to grow out of categories and concepts that initially emerged from
this textual analysis.
The burden of this analysis is how aptly the Collins, Evans, and Gorman trading
zone model describes the evolution of the trading zone found in the TED case study.
Thus, in order to clearly present this research, I have integrated the analysis of trading
zones within the presentation of the case study.
Background
In their 2007 paper, Collins, Evans, and Gorman presented a model that defines the
space through which trading zones can evolve (figure 8.1). Defined by the parameters
of homogeneity, heterogeneity, collaboration, and coercion, the model describes four
ideal trading zone types: enforced, fractionated, interlanguage, and subversive. An
enforced trading zone can occur when there are high levels of coercion and hetero-
geneity. The authors give the example of slave labor, in which the slavers and the
enslaved are distinctly different groups, and the trade of service for food and relief
from punishment is maintained by physical force. A fractionated trading zone is
highly collaborative and highly heterogeneous. Within this type of trading zone are
two subcategories: boundary object trading zones and interactional expertise trading
zones. In boundary object trading zones, the operative medium is a physical item—
which may hold different meaning to the parties involved in the trade—rather than
The Evolution of a Trading Zone 159
Figure 8.1
Parameters of trading zones.
a linguistic exchange. For instance, cowrie shells may be of monetary value to one
group of people or of scientific value to another. In contrast, interactional expertise
trading zones are mediated by language mostly in the absence of a physical item or
items. An interlanguage trading zone forms in situations that are highly homogeneous
and highly collaborative. In this type of trading zone, new fields, such as biochemistry
and nanoscience, can emerge from the joining of existing fields. Finally, a subversive
trading zone is highly homogeneous and coercive. An example of this type of trading
zone would be the prevalence of Microsoft Windows-dependent software; although
some people might prefer to use other operating systems, they have no choice if they
want to use this software.
Case Study
For many years, scientists and managers had been aware of most causes of sea turtle
mortality, but they did not know the relative contribution of each cause to sea turtle
population decline. However, it was commonly accepted in the scientific community
that a major source of mortality for juvenile and adult turtles was drowning in shrimp
trawls. The assumption was based on the increased number of turtle carcasses observed
during a yearly time of overlap between high concentrations of breeding sea turtles
during nesting season and extensive shrimp trawling near nesting beaches during
shrimping season. Shrimp and sea turtles share the same habitat—coastal waters along
the southeastern United States. Shrimpers tow large nets known as trawls for two to
three hours at a time. Sea turtles encountering these nets often attempt to escape, but
some are captured. Unable to surface to breathe, many turtles drown during these
long tow times (NRC 1990). Many members of the shrimping industry did not accept
the claim that bycatch was adversely affecting sea turtle populations, in part because
individual shrimpers may only capture one or two turtles each year. It was difficult
for many of them to appreciate the cumulative impact that tens of thousands of
shrimpers each catching one or two turtles could have on sea turtle populations (NRC
1990).
Sea turtle bycatch became a management issue for the shrimp fishery after the
listing of several sea turtle species under the Endangered Species Act. During the 1970s,
the National Marine Fisheries Service (NMFS) listed as threatened or endangered all
six species of sea turtles that occur in U.S. waters. NMFS, which manages marine fish-
eries, was also responsible for protecting sea turtles. In order to fulfill this responsibil-
ity, NMFS began research to invent a device to reduce sea turtle bycatch.
Enforced Trading Zone: Pressured to Collaborate
In 1976, NMFS began research and public education programs to reduce sea turtle
mortality in shrimp trawls by creating modifications to fishing gear. At the beginning
of its research program, NMFS consulted with two sea turtle specialists, one a govern-
ment employee and one an academic, about initial designs for gear modifications.
NMFS did not consult with shrimpers, even though they were given evidence that
shrimpers might have relevant expertise to share. Specifically, in 1976, Sea Grant (the
agency charged with transferring knowledge and technology to the fishing industry)
notified NMFS of a device, called a jellyball shooter, which some shrimpers were using
to remove large jellyfish and marine debris from their nets. Sea Grant suggested that
this same approach could be applied to removing sea turtles from a net. Ignoring the
jellyball shooter idea, NMFS pursued the line of invention suggested by the sea turtle
specialists, which was a barrier panel to prevent turtle capture by barring their entrance
to the net.
In 1977, members of the shrimping industry became involved in addressing the
issue of turtle bycatch. This nascent trading zone originated when NMFS began to
study the extent of the turtle bycatch problem via a cooperative observer program
with shrimpers. In this same year, the president of the Southeastern Fisheries Asso-
ciation invited an NMFS scientist to speak about conceptual ideas for preventing
turtle bycatch at the association’s annual meeting. Association members chuckled
and in general refused to believe that turtle bycatch was such an extensive and
important issue. Subsequently, the membership of the Texas Shrimp Association and
affiliate associations had a similar reaction to this issue. Yet, in contrast to their
members, the leaders of these organizations recognized the large potential impact
of turtle bycatch on the shrimp industry and spent years trying to rally the shrimp-
ing industry to be proactive about turtle bycatch. For its part, NMFS made
only cursory requests for collaboration. NMFS felt the pressure to move forward
quickly to produce a solution to the sea turtle bycatch problem. It is possible that
NMFS’s cursory invitations were a necessary courtesy to its major stakeholder in the
region—the shrimping industry—but in actuality NMFS was content to pursue only
its own research goals.
At this point (figure 8.2, stage 1), shrimpers and government scientists remained
heterogeneous rather than homogeneous and were being encouraged to collaborate.
Neither group, but especially the shrimpers, felt a significant need for this collabora-
tion, so the trading zone—to the extent that it existed in the form of the cooperative
observer program—was coerced rather than collaborative at this stage. The coercive
force for the shrimpers was the pressure from their industry leaders. As for NMFS, it
had a need to collect data coupled with a need to maintain relations with the shrimp-
ing industry. Both these needs were motivated to a large extent by top-down pressure
within NMFS to begin addressing the sea turtle bycatch problem. Thus, this stage, with
its coercive pressures, is an example of an enforced trading zone.
Enforced Trading Zone: Institutional Power
Fieldwork on the barrier panel began in 1978 with limited collaboration with two
fishing industry organizations (figure 8.3). For the first two years of the research
program, NMFS only tested gear ideas that it had originated. Initially the collaboration
with industry was limited to the use of commercial vessels to conduct sea turtle popu-
lation studies and to test prototype fishing gear. The shrimpers served only as captain
and crew for these vessels, while NMFS personnel oversaw the experiments.
The close proximity of the two groups aboard the vessels led to exchanges about
the experimental gear. Many of the shrimpers felt that they had relevant expertise to
share about how to improve the gear’s performance. Some shrimpers who were
involved in the cooperative program felt that their ideas were ignored or belittled in
1-pressure to collaborate
Need for controlled

use of shrimper
expertise
2-institutional power
Figure 8.2
Evolution of TED trading zones.
Figure 8.3
Barrier panel (Watson 1980).
comparison to the ideas of NMFS personnel. One shrimper who felt that his ideas were
well received by NMFS stated that he believed his experience was atypical. The shrimp-
ers’ perception was that NMFS personnel were considered experts in experimental gear
modification and shrimpers were not. NMFS’s resistance to the shrimpers’ ideas may
have been a result of the agency’s attempt to limit experimental variables, as stated
in an NMFS report:
Although the use of cooperative commercial shrimp vessels for testing excluder gear is essential
to the successful development of an operational excluder trawl, an accurate estimate of trawl
performance, particularly shrimp loss estimates, is difficult on vessels where full control cannot
be exercised. Sources of error are introduced into the experimental design using a cooperative
commercial fishing operation. In order to minimize the sources of error, the number of coopera-
tive vessels must be kept at a level where the number of available qualified gear technologists
can closely monitor each vessel’s operation during the experiment. (Watson 1980)
Here, the fishers are essentially identified as sources of error whose practices should
be routinized and controlled to make the development of TEDs possible. Yet the
quotation also indicates that the trading zone had evolved. While still an enforced
trading zone, it is at this point more characteristic of “institutional power” (figure
8.2, stage 2). This type of enforced trading zone exists when the recognized exper-
tise of an established discipline is such that it has the authority to decree whether
or not certain knowledge belongs within that discipline or group (Collins, Evans,
and Gorman 2007). In essence, “the establishment” polices its borders to insure
that integrity of the discipline is not weakened by knowledge that it judges to be
lacking.
The transition from pressured collaboration to institutional power in this case was
fueled by increased proximity between the two groups and the desire for NMFS to test
its designs in real-life fishing situations. This desire came from the NMFS field-level
scientists rather than higher-level policymakers. It created a reason more compelling
than stakeholder relations to engage in a trading zone with industry.
In this top-down state of “institutional power,” NMFS orchestrated the invention
and development process and determined the involvement of all other players. NMFS
determined how the invention and development process should proceed, as is evi-
denced by the report quoted above (Watson 1980). Furthermore, NMFS or one of its
agents, such as Sea Grant or a state agency, to which NMFS delegated this responsibil-
ity solicited and selected the specific fishers whose vessels were used in the cooperative
tests. In this way, NMFS pigeonholed these fishers into specific roles, while failing to
perceive the range of other possible roles that the fishers could have played in the
invention and development process.
In this new stage, NMFS held most of the power. It had the financial resources to
fund large-scale research, equipment resources—such as underwater cameras—to
conduct the research, and personnel resources to analyze the results and make further
gear modifications. Although it would have been feasible for NMFS to conduct the
gear modification research without any assistance from the shrimp industry, the
industry’s involvement allowed them to test devices under a larger range of real-world
circumstances. From this position of power, NMFS could and did define what the
relationship would be between it and its industry collaborators. NMFS had the power
to engage only collaborators that would cooperate under their terms. One of these
terms was that NMFS would monitor the activity of its industry collaborators and try
to maintain full control over the vessels’ operation. Shrimpers willingly engaged in
this unbalanced trade because they were being financially compensated.5 This tradeoff
was also the only way at that time that they could engage in the government’s gear
invention process to reduce sea turtle bycatch.
Fractionated Trading Zone: Boundary Object
In 1980, after two years of field tests, major problems persisted with the barrier panel.
Namely, the device resulted in high shrimp loss rates and was not adaptable enough
for commercial use. That same year, Sea Grant sent NMFS photos of a “jellyball
shooter” and again suggested that a similar approach could work for excluding turtles.
The jellyball shooter had been used for decades, especially by shrimpers in South
Carolina and Georgia, when cannonball jellyfish were so dense that shrimping could
not otherwise occur. The jellyball shooter consists of a grid similar to a charcoal grill
that is placed in the neck of the trawl to block large objects from entering the net bag
and directs such objects out of a hole cut in the net. Shrimp are small enough to be
carried by the water flow between the bars of the grid and into the end of the net.
This time the jellyball shooter sparked interest within NMFS. Based on the photo-
graphs, independently and simultaneously, an NMFS scientist and a contract vessel
captain conceived of placing the grid within a frame. The NMFS scientist, working
from his scientific paradigm of hydrodynamics and net dynamics, constructed his
version from light PVC. The shrimping captain, working from his fisher’s paradigm
and being familiar with the severe stress and wear that commercial fishing can put on
gear, constructed his from sturdy steel. Though the shrimper’s original model was too
heavy and the scientist’s too fragile to be practical, NMFS drew ideas from both to
apply to a new design. NMFS called the resulting prototype the turtle excluder device
or TED (figure 8.4).
Figure 8.4
NMFS TED.
The use of the jellyball shooter photos and TED prototypes as a means to exchange
ideas between groups is indicative of a boundary object trading zone (figure 8.5, stage
3). In this type of trading zone, exchanges between parties occur through a material
medium (e.g., a physical item), largely without linguistic interchange. Often the
participants in this type of trading zone interpret the meaning and significance of
the object differently, perhaps even obscuring the role of other parties. However,
products that the exchange yields are valuable enough to keep the parties engaged
in the trade.
This transition from a trading zone enforced by institutional power to a boundary
object fractionated trading zone was compelled by the failure of the barrier panel,
which had such high shrimp loss rates that it was commercially impractical. The
expertise within the institution of government gear research proved insufficient to
solve the design flaws of the barrier panel. Furthermore, the increasing political pres-
sure to reduce sea turtle bycatch in compliance with the Endangered Species Act made
NMFS more willing to consider ideas from outside the institution.
However, NMFS experts and shrimp fishers operated from different paradigms,
hindering direct communication between the two groups. Scientists and shrimpers
did not share the same superordinate goal regarding TEDs. Because scientists viewed
TEDs as a means to protect and conserve sea turtles, most of their research focused
primarily on excluding turtles from the nets. The practicality of the device for com-
mercial use was a more distant second priority, as is illustrated by the government’s
testing protocols to certify a TED for legal use. While the certification standard for
4a-interactional expertise
Sea Grant
engages
3-boundary object
Institution expertise insufficient
Need for controlled

use of shrimper
expertise
2-institutional power
Figure 8.5
TEDs was based on a turtle exclusion rate of 97 percent, there was no mandated shrimp
retention rate that TEDs had to satisfy.
Shrimpers, on the other hand, viewed TEDs as a means to fulfill the letter of the
law in order to protect their privilege to continue fishing profitably. Most shrimper
innovations focused on increasing shrimp retention and the usability of TEDs under
various commercial fishing conditions. These innovations included TEDs that were
less likely to clog with weeds, TEDs made from soft material that would be more pliable
in fishing gear, and roller devices to allow the TED to move easily over a rough
seafloor.
The two communities also differed on what they regarded as data. As evidenced by
the previous excerpt from a 1980 NMFS report (Watson 1980), NMFS strove to achieve
better-controlled experiments by reducing variables. For many shrimpers, their
personal experience of using a device while fishing provided adequate data on which
to make a decision of how best to improve the device. For example, there is a long-
standing debate between members of the science and shrimping communities about
the best way to test TEDs. Many of the scientists support an experiment that uses
captive-reared sea turtles to test TEDs under controlled conditions. As a result of these
tests, the scientists believe that TEDs that open to the top will release turtles more
effectively. Notably, NMFS raised these captive sea turtles in shallow tanks that may
have affected their diving/escape response. Members of the shrimping community
argue that their years of experience on the water have shown that wild turtles will
attempt to dive down to escape harm, not swim up. Thus, they believe that the
captive-reared sea turtles have an unnatural response and that bottom-opening TEDs
will release sea turtles more effectively. Furthermore, as evidence that shrimpers’
superordinate goal regarding TEDs is to be able to fish legally and profitably, the
shrimpers point out that a bottom-opening TED would also passively release debris,
such as plastic and logs, that could otherwise clog the TED and cause shrimp loss or
crush the shrimp and reduce its value.
In the years following the invention of the NMFS TED, NMFS refined and improved
the TED, and individual shrimpers made their own modifications to the TED outside
of the government invention system. The critical role that the jellyball shooter pho-
tographs and first prototype TED designs played in the creation of the NMFS TED set
the stage for many future exchanges between government scientists and shrimpers.
There is limited evidence in the historical record of direct communications between
government scientists and those members of the shrimping industry who were actively
working to invent and modify TEDs. The primary means of exchange between the
two groups about TED modifications were prototypes and representations of the
devices, such as schematics, photographs, and videos. These representations were also
accompanied by written descriptions of the device. Representations of various TED
designs spread virally throughout the southeastern United States via photocopies,
dubbed videos, and promotional pamphlets. As a result, people across eight states who
most likely had never directly communicated were able to create slightly modified
copycat devices, based on the best-known designs.
This disassociated relationship indicates a boundary object trading zone, which is
well illustrated by the development of soft TEDs. Soft TEDs have an exclusion panel
made of net rather than a hard grid. The inventor of the first soft TED was inspired
by diagrams of fishing gear used on the West Coast of the United States that he had
seen in a national fishing magazine. Although the inventor was a soft-spoken fisher-
man and did not travel extensively to promote his device, soft TEDs quickly spread
throughout the southeastern United States. After NMFS personnel tested and certified
the device for legal use, they and Sea Grant distributed promotional materials com-
plete with diagrams of the soft TED. Vessels using the device brought it with them
when fishing in the waters of other states. In the span of about five years, at least
thirteen unique types of soft TEDs existed. Each of these was created by different
inventors who mostly came from within the fishing, net-making, and gear-manufac-
turing communities. These inventors represented seven different states, and many of
the designs had multiple versions. The evolution of these devices over time clearly
shows the influence of previous designs; however, with the exception of one case,
there is no evidence of extensive direct communication between the inventors. This
suggests that the soft TEDs served as a boundary object through which to exchange
information.
Fractionated Trading Zone: Interactional Expertise
Even though the original TED incorporated the ideas of both fishers and government
scientists, much of the shrimping industry viewed it with suspicion, because scientists
were working within the government-controlled invention system. The shrimping
industry felt that government personnel had developed the TED in a government
backroom and were forcing it on them as the only acceptable solution to the turtle
bycatch problem.
Because of the controversial nature of this situation, Sea Grant had kept its distance.
Sea Grant engages with the fishing community on numerous issues, such as introduc-
ing new fishing technologies and educating the industry about new regulations. Most
often Sea Grant serves to transfer information from NMFS to the fishing industry, and
thus Sea Grant agents must speak the language of government scientists and managers
as well as the language of fishers. As a result, they are, in essence, professional inter-
actional experts; and this interactional expertise could have served to broker more
productive trades between NMFS and the shrimping industry. In other words, Sea
Grant was in a position to play the role of a boundary organization, in that it could
provide the opportunity for the creation and use of boundary objects (here, the TEDs)
and could mediate between and draw in the participation of players from both sides
of the boundary (Guston 2001). Unfortunately, given the range of its responsibilities,
Sea Grant was wary of spending precious social capital on the controversial sea turtle
issue, especially in light of NMFS’s controlling approach to invention and its previous
resistance to the fishers’ and Sea Grant’s ideas.
In 1982, in order to overcome Sea Grant’s reticence, an NMFS manager staged an
exchange to secure Sea Grant’s active support in TED development and promotion.
The NMFS manager asked shrimping industry representatives to publicly request Sea
Grant’s help with technology transfer during a workshop. The industry representatives
then assured Sea Grant that they wanted their assistance with the TED issue, and
helped ease Sea Grant concerns that the controversial nature of TEDs might tarnish
its relationship with the industry.
In this way, the interactional expertise of Sea Grant was brought to bear, and the
trading zone transitioned from a boundary object trading zone to one of interactional
expertise (figure 8.5, stage 4a). Unlike the previous transitions, which were rather
abrupt, the transition to an interactional expertise trading zone occurred over a
number of years. This slower transitional time is likely a common characteristic of
interactional expertise trading zones. Unlike some of the other trading zones in which,
as I have shown, transitions resulted from a change in circumstances, interactional
expertise involves the time-consuming development of a new linguistic ability. Even
in the case of Sea Grant agents, who already had this ability, their use of it was depen-
dent on establishing relationships between NMFS and shrimpers through which to
exercise this linguistic ability.
In the TED case, Sea Grant agents developed these relationships over the course
of four years. They engaged with shrimpers who had begun to develop new TEDs
on their own, outside of the NMFS invention system. They also interacted with
NMFS about the continued refinement of the NMFS TED. Sea Grant agents recog-
nized that, in order to get shrimpers to accept TEDs, the devices would need to be
more practical for commercial use. Furthermore, they believed the best way to
achieve this would be to bring shrimpers’ TED ideas to the attention of NMFS
scientists. So in 1986, Sea Grant sponsored a demonstration event comparing the
NMFS TED with three shrimper-invented TEDs. Drawing on their interactional exper-
tise and cultural understanding of the two groups, Sea Grant convinced shrimpers
and government scientists to participate. As a result of this event and further testing,
NMFS certified all three shrimper-invented TEDs for commercial use. This event
firmly established the interactional expertise trading zone, which continues to this
day. Innovative shrimpers continue to bring their ideas for new TEDs to Sea Grant
agents, who then communicate these ideas to NMFS. Likewise, NMFS still depends
on Sea Grant to help communicate the results of TED testing and ideas of how to
improve prototype TEDs.
Divergence of Trading Zones
Upon the creation of the first NMFS TED, there was increasing public and political
pressure to promote this device for commercial use. This new objective of promoting
the adoption of TEDs was substantially different from the previous and ongoing
4a-interactional expertise
4b-boundary organization
Sea Grant
engages New objective
3-boundary object
Trade not
meeting
parties’ needs
Institution expertise insufficient
Need for controlled

use of shrimper
5b-mediation expertise
Structure
inadequate 2-institutional power
for trade
6b-regulatory mandate
Figure 8.6
objective to develop a device to reduce turtle bycatch in shrimp trawls. Because these
objectives were so different, the existing boundary object trading zone (figure 8.5,
stage 3) could not adequately serve to address the new goal. For this reason, a new
boundary organization trading zone diverged from the boundary object trading zone.
This new trading zone involved many of the same people, but the nature of their
engagement changed significantly with the new objective and the new boundary
organization, called the TED Voluntary Use Committee (figure 8.6, stage 4b).
An environmental nonprofit group initiated the formation of the TED Voluntary
Use Committee to help promote TED use. This egalitarian committee consisted of
representatives from shrimp industry and conservation organizations as well as NMFS
and Sea Grant, and was cochaired by an environmentalist and a shrimping industry
leader. In 1983, the committee agreed upon a goal that within three years the majority
of southeastern U.S. shrimpers should be using TEDs; areas and times of critical impor-
tance to sea turtles should be identified as soon as possible; and that usage of TEDs
should be 100 percent in areas of critical importance, especially areas frequented by
the highly endangered Kemp’s Ridley sea turtle (Weber 1995).
In the early stages of the Voluntary Use Committee, all the participating parties
strove to achieve these goals. Often acting autonomously, the parties conducted
studies, published educational literature, and held TED workshops and demonstra-
tions in an effort to promote TED adoption. The nature of this engagement was
characteristic of a boundary organization, in that the Voluntary Use Committee and
its goals were the common elements around which the parties interacted. These goals
served as the common motivation for their activities. But, as is also characteristic of
a boundary organization trading zone, the parties had additional external motivations
and goals not shared or understood by the other parties. This led to different inter-
pretations of how to achieve the common goals and a limited understanding of the
reasoning behind these interpretations by the other parties. Despite these differences,
this trading zone can function as long as the activities of the parties continue to con-
tribute to what all parties view as common goals.
Several years into the voluntary use program, it became apparent that the program
was not fulfilling the committee’s goals. In 1983, NMFS estimated that only 150 vessels
out of tens of thousands were using TEDs at least part-time (Weber 1995). In the same
year, NMFS surveyed eighty TED recipients. While the majority of the thirty respon-
dents said that they had used and would continue to use the TEDs, they also admitted
that they were likely to use them only in debris-filled waters (Jenkins 2006, 652). By
1985, Louisiana Sea Grant estimated that 300 shrimpers were using some type of
excluder or cannonball shooter to exclude jellyfish in Cameron County. Although
there were no comprehensive surveys, estimates based on TED giveaways, orders
placed to manufacturers, and Sea Grant observations of homemade TED use were
disappointing. Five years into the voluntary use program, at best 5 percent—but more
likely far less—of shrimpers were using any type of excluder device even part-time
(Weber 1995).
In light of the low levels of voluntary use, the cooperative spirit between parties
began to fail. In November 1985, the TED Voluntary Use Committee held a meeting
to evaluate the program’s progress and to discuss additional efforts to promote volun-
tary TED use. The industry representatives and environmentalists could not agree on
what the future TED adoption goal for the program should be (Mast 1985). During
what was perhaps the last meeting of the Voluntary Use Committee, frustration levels
reached a critical point between environmentalists, industry members, scientists, and
government. NMFS sought to prolong the voluntary use program for another two to
four years, because they had created a commercially practical TED only in 1984. The
industry also sought to prolong voluntary use for this same reason, but also to buy
time for shrimpers to work without TED requirements. The scientists from the Sea
Turtle Recovery Team advocated the dual pursuit of voluntary use and development
of mandatory TED use regulations, if they proved necessary in the future. The envi-
ronmentalists pushed for a firm commitment from the industry group leaders to
persuade a certain percentage of their membership to adopt TEDs by a set date. Neither
of the shrimping industry leaders on the Voluntary Use Committee was willing to
make this commitment.
At this point, what the environmentalists and industry leaders agree had been an
amicable, sincere, and respectful working relationship began to sour. The environmen-
talists believed that their fellow committee members from the shrimping industry
should use their positions as the preeminent industry leaders to sway their member-
ships. The industry leaders, however, viewed their role on the committee more as
being representatives for their members’ views, facilitators of NMFS and Sea Grant
extension efforts toward their members, and encouragers of the use of TEDs, but cer-
tainly not as persuaders for their use. They realized that their positions as directors of
their organizations (the Texas Shrimp Association and Southeastern Fisheries Associa-
tion) hinged on their ability to negotiate the best deal on issues that affected their
members, while remaining within the bounds of what their members could tolerate.
Being forced to use TEDs was definitely outside of these bounds. In fact, the tenure
of one committee member as the director of the Texas Shrimp Association ended in
part because members of the association viewed him as too pro-TED. (Researchers have
observed similar tensions in the field of AIDS research: as AIDS activists became full-
fledged experts and partners in the scientific process of drug development, their
expertise caused a growing divide between them and their fellow activists who had
not developed scientific expertise [Epstein 1995].)
The closing of 1985 marked the effective, if not official, end of the voluntary use
program. This boundary organization trading zone had failed because of the external
forces on the parties and the participants’ lack of understanding of these forces. For
each party, the importance of external goals had increased, resulting in changes in
how they interpreted the common goals. Unfortunately, these interpretations were
increasingly conflicting. Also, the environmentalists believed that the shrimping
industry leaders were not contributing as much as they could to the common goal of
high levels of TED adoption. In reality, the shrimping leaders were contributing as
much as they could to the process, given the external constraints placed on them by
the members of their organizations. Because boundary organizations trading zones

tend to obscure a large portion of the external roles and activities of other parties, the
environmentalists lacked understanding of the external constraints on the shrimping
industry leaders and thus believed that the industry leaders were “holding out” on
the trade. It is also possible that the trading zone was too small to achieve its goals.
The shrimping community consisted of tens of thousands of diverse shrimpers, and
many of them did not belong to an industry group. These nonaffiliated shrimpers
may not have been easily influenced by industry leaders.
Having reached an impasse, the environmentalist cochair of the Voluntary Use
Committee wrote the director of NMFS’s Southeast Regional Office to suggest that
NMFS implement mandatory TED use regulations. When NMFS failed to act on this
request, a leading environmental organization informed the Secretary of Commerce
in August 1986 that it intended to sue the Department of Commerce over failure to
enforce the Endangered Species Act (Street 1987; Weber 1995). Fearing that NMFS
would close the shrimp fishery in response to the threatened lawsuit, the Texas Shrimp
Association asked NMFS to sponsor mediation between the industry and the environ-
mental community (figure 8.6, stage 5b).
The mediation team that NMFS put together consisted of six environmentalists,
representing four groups, and six industry representatives from six different industry
groups. The Administrator of the National Oceanic and Atmospheric Administration
(the parent agency of NMFS) served as the mediator between the environmentalists
and shrimping industry representatives. Participants in the mediation recall that
the Administrator took a heavy-handed approach. In a separate room, he lectured the
shrimping industry representatives, emphasizing the direness of the situation and the
potential for either mandatory TED use by all shrimpers or a complete shutdown of
the fishery if the shrimping representatives did not make a deal. The shrimping rep-
resentatives perceived this as a threat (Margavio and Forsyth 1996). The coercive
nature of the interaction between the mediator and the shrimping industry representa-
tives is indicative of an enforced trading zone.
After several meetings in 1986, the mediation team reached a compromise that
would require TEDs in “hotspots,” areas where turtles were in high concentrations
during certain times of the year. The agreement was signed by all the parties except
for one industry representative, who promised to sign later. Subsequently, this repre-
sentative refused to sign the agreement. Furthermore, the Texas Shrimp Association,
whose representative had signed the agreement, repudiated it, claiming that the orga-
nization had not given him the authority to sign. As a result, the other industry groups
also withdrew from the agreement (Margavio and Forsyth 1996).
Stakeholders in the TED case generally view the mediation as a failure, placing
much of the blame on the industry representative who refused to sign. However,
considering the diversity of the constituents being represented, it is extraordinary that
the mediation team was able to draft and verbally consent to the agreement. While
the environmentalists presented a rather uniform platform, the six industry members
were intended to represent tens of thousands of shrimpers from eight different states.
These shrimpers varied from one-man inshore operations to owners of offshore fleets.
The needs and desires of these shrimpers varied inasmuch as their shrimping opera-
tions varied. Given this variation, it is truly astounding that the mediation proceeded
as far as it did. Even if an agreement had been reached, many of the key stakeholders
in the TED case believed it was unlikely that the large and diverse body of shrimpers
would have felt they had been adequately represented by the six industry representa-
tives at the mediation table. Given this, it is probable that many shrimpers would not
have abided by the mediation agreement.
This consideration suggests that the structure and nature of the trading zone was
inadequate for the type of trade that needed to occur. If the participants had applied
the concept of trading zones and been conscious of the state of their trading zone,
they might have recognized the inadequacy of mediation for achieving their goal and
might have pursued another type of trading zone. The concept of trading zones can
be of practical use in real-world problem solving and decision making.
When the mediation failed, NMFS promulgated mandatory TED use regulations in
1987, requiring the use of TEDs by all shrimpers in all waters at all times. After years
of amendments to the regulations and varying levels of enforcement, the regulations
became fully and consistently in effect in 1994 (Weber 1995). The penalties for failing
to comply included the seizure of a vessel’s shrimp catch and large fines. The regula-
tions created a new enforced trading zone further down the coercive scale, which still
remains in place (figure 8.6, stage 6b).
Adapting the Trading Zone Model
In this paper, I have presented the case study of the development and use of TEDs
and the types of interactions among stakeholder groups that occurred as a result of
the effort to reduce sea turtle bycatch. The trading zones concept offered a frame for
understanding why these trading zones began, changed, and failed. In general, I found
that the forces that drove these transitions were either the development of new goals
or the inadequacy of the trade to meet existing goals (or, in some cases, both factors).
Often the transition between trading zones can be abrupt, resulting in a change in
circumstances. But in the case of interactional expertise, the transition is longer,
because a new linguistic skill must be learned and more intimate relationships
developed.
The exercise of applying the trading zone model proposed by Collins, Evans, and
Gorman (2007) to the TED case study yielded useful insights into the nature of
exchanges between stakeholders. I have shown how the concepts of fractionated and
enforced trading zones were useful for describing these engagements. Consistencies
in the characteristics of the specific trading zones that fell into these larger categories
of fractionated and enforced indicate the normative value of these categories. Notably,
the only dynamic parameter in the TED case study was the level of collaboration/
coercion, so the parameter of heterogeneity remains mostly unexplored. The one
instance of significant movement along the heterogeneity scale was in the interac-
tional expertise trading zone (figure 8.5, stage 4a). In this case, the heterogeneity of
the two groups—shrimpers and scientists—changed little. But the addition of a third
group—Sea Grant agents—who possessed characteristics of both shrimpers and
scientists, including the ability to communicate with both groups, smoothed the
continuum between the groups, making all the players taken en masse appear more
homogeneous.
Upon structuring the case study within the Collins, Evans, and Gorman trading
zone model, I found most of the specific trading zones to be of only descriptive
value—specifically, the pressure to collaborate, institutional power, mediation, regu-
latory mandate, some of which are terms borrowed from those authors and others
of my own creation. The distinctions between these trading zones were often
minimal and situational, with no clear boundaries of what delineated each. The
exceptions to this observation were the boundary object, boundary organization,
and interactional expertise trading zones. These specific trading zones had a clear
separation delineated by the level of linguistic exchange and the presence of a
material culture. This may suggest that there are subcategories within the larger
trading zone category. For example, using the Collins, Evans, and Gorman model,
the subcategory of interactional expertise may be bound by high collaboration and
low heterogeneity (without slipping into the realm of homogeneity). Likewise, the
subcategory of boundary object may be defined by low collaboration (without cross-
ing into coercion) and high heterogeneity. In essence, this boundary would diago-
nally bisect the “fractionated” quadrant in the Collins, Evans, and Gorman model;
however, later in this paper I will describe an adapted model that I believe gives
more intuitive parameters for bounding the subcategories of interactional expertise

and boundary object.
Upon structuring the TED case study within the Collins, Evans, and Gorman model,
I found that conceptually, if not empirically, there were several aspects of the TED
case study that did not fit well with the model. Collins et al. intended the model as
a simple four-by-four chart. However, in order to illustrate the moving evolution of a
trading zone, I interpreted the model as a scaled diagram. Thus, the heterogeneity/
homogeneity axis suggests but does not actually have a null point; given this, there
would not logically be a negative side of the heterogeneity/homogeneity axis. Second,
as presented, the model suggests that all states of being are equally likely to produce
a trading zone. This is appropriate for Collins et al.’s purpose of defining the trading
zone categories. However, as I have applied it, a model that can also be used to chart
the evolution of a trading zone would serve better if it depicted the conditions in
which trading zones were more likely to form.
Third, the position of collaboration and heterogeneity at the top right of the model
places these parameters in what would be positive positions on a graph. This location
inadvertently places a positive value on parameters that are already deemed
favorable—as a gross generalization—by societal perceptions. Because one of the
benefits of the trading zone concept is to describe the role and appropriateness of a
trading zone for different situations, it is important to present each trading zone type
as equal and neutral in terms of value.
Finally, while heterogeneity/homogeneity is an important characteristic of trading
zones, it was not a defining parameter for the TED case study. Rather, I hypothesize
that changes in heterogeneity/homogeneity are driven by another variable, such as
the degree of shared language. It stands to reason that, as the ability to communicate
increases, the flow of information that can result in changes of another culture will
also increase, thus altering the degree of heterogeneity.
Taking all of these observations into consideration, I offer an adapted trading zones
model (figure 8.7). I have turned the model on its side and switched the polarity of
the collaboration/coercion scale, so as to mitigate the positive value usually associated
with the right end of the scale. The heterogeneity/homogeneity axis is now the
“shared language” axis, with a null point of no shared language. The area that is not
as likely to originate a trading zone, such as independence coupled with little shared
language, is denoted by shading. Although I believe this area is likely not to originate
a trading zone, it would be possible for an evolving trading zone to move through
this area. However, it is reasonable to hypothesize that the trading zone is more likely
to fail while in this area.
SHARED LANGUAGE
INTERLANGUAGE SUBVERSIVE
threshold threshold
FRACTIONATED
Interactional
expertise
Unfavorable conditions for a
trading zone to originate
ENFORCED
FRACTIONATED
Boundary
object/organization
COLLABORATION COERCION
Independence and
no shared language
Figure 8.7
Adapted trading zone model.
The boundary between this unfavorable area and trading zone areas is demarcated
by a curved line. This curved line illustrates the hypothesis that there is a threshold
at which a level of shared language is obtained that allows the rapid accumulation of
additional shared language. It is important to note that even though this model
appears as a graphical representation, I do not intend to imply a dependent relation-
ship between the two axes.
Each trading zone category is labeled, including the subcategories of interactional
expertise and boundary object/organization. The relative areas of the spaces in the
model have no significance, with the exception that the smaller areas for interlanguage
and subversion were intentional. As the TED case study indicates, obtaining the state
of interactional expertise can be effortful, because of the time needed to acquire
language skills and foster relationships. Arguably, then, as one moves further up the
language axis, more effort and time may be involved, thus decreasing the likelihood
of obtaining language-dependent trading zones, such as interlanguage and subversion.
With the substitution of shared language for heterogeneity/homogeneity, the defini-
tions for the trading zone categories presented by Collins, Evans, and Gorman still
hold true because of the close tie between common language and homogenization.
Structuring the TED case within this new adapted trading zone model shows that
the pattern of evolution is much the same, again due to the close relationship between
SHARED LANGUAGE
INTERLANGUAGE SUBVERSIVE
FRACTIONATED ENFORCED
Interactional
expertise
4a
5b
1
FRACTIONATED
4b 3 2
Boundary
6b
object/organization
COLLABORATION COERCION
Independence and
no shared language
Figure 8.8
TEDs and the adapted trading zone model.
shared language and heterogeneity/homogeneity (figure 8.8). However, the new axis
better speaks to the variable that actually changed—linguistic exchange—in the TED
case study. In future work, the parameters of both this and the Collins, Evans, and
Gorman model should be tested by applying other trading zone case studies, especially
those of subversive and interlanguage trading zones. Such repeated applications should
yield more insight into the defining parameters of trading zones. Such a study would
also prove the existence of an area with a low probability of producing trading zones,
and would provide its dimensions.
Conclusion
It is clear from the TED case study that the concepts of enforced and fractionated
trading zones hold practical utility for describing and defining real-world exchanges.
I have presented evidence that trading zones do indeed evolve, and have identified
drivers for this evolution. Furthermore, I illustrate the divergence of two trading zones.
Finally, I have presented an adapted trading zones model that is conceptually, if not
empirically, a better fit for the TED case study. In order to determine the normative
value of all the trading zone categories as well as to continue to explore trading zone
models, it is imperative that numerous case studies be applied to test their fitness.
Notes
1. I chose not to include this more recent TED history in the case study because the circumstances
surrounding it differ greatly from those of the previous regulations.
2. The study considers individuals as representatives of organizations; verbal and written state-
ments are taken to be valid representations of the opinions, attitudes, and beliefs that guided
the work of those organizations.
3. To learn more about the socioeconomic aspect, policy, or other aspects of this case study,
please see Margavio and Forsyth 1996; Maril 1983; Maril 1995; Durrenberger 1996; Maiolo 2004;
and NRC 1990.
4. For more on snowball sampling, see Bernard 2002.
5. It is a common practice for NMFS to compensate industry collaborators financially by either

paying them directly, increasing the allowed catch limit, or allowing fishing during a time or
in an area that is typically closed. But in this case, because of the nature of the shrimp fishery,
direct payment is the sole option that is commensurate with the time, effort, and resources that
shrimpers expend in these research activities.
References
Bernard, H. R. 2002. Research Methods in Anthropology: Qualitative and Quantitative Approaches.

Walnut Creek, CA: Altamira Press.
in History and Philosophy of Science 38:657–666.
Durrenberger, E. P. 1996. Gulf Coast Soundings: People and Policy in the Mississippi Shrimp Industry.
Lawrence: University Press of Kansas.
Epstein, S. 1995. The Construction of Lay Expertise: AIDS Activism and the Forging of
Credibility in the Reform of Clinical Trials. Science, Technology and Human Values 20 (4):
408–437.
Guston, D. H. 2001. Boundary Organizations in Environmental Policy and Science: An Introduc-

tion. Science, Technology and Human Values 26 (4):399–408.
Jenkins, L. D. 2006. The Invention and Adoption of Conservation Technology to Successfully

Reduce Bycatch of Protected Marine Species. Ph.D. diss., Nicholas School of the Environment
and Earth Sciences, Duke University.
Maiolo, J. R. 2004. Hard Times and a Nickel a Bucket: Struggle and Survival in North Carolina’s Shrimp
Industry. Chapel Hill, NC: Chapel Hill Press.
Margavio, A. V., and C. J. Forsyth. 1996. Caught in the Net: The Conflict between Shrimpers and
Conservationists. College Station: Texas A&M University Press.
Maril, R. L. 1983. Texas Shrimpers: Community, Capitalism, and the Sea. College Station: Texas A&M
University Press.
Maril, R. L. 1995. Bay Shrimpers of Texas: Rural Fishermen in a Global Economy. Lawrence: University
of Kansas Press.
Mast, R. 1985. Status of the Sea Turtle Rescue Project and Progress in the Use of the TED. Kenner, LA:
NMFS Trawling Efficiency Device Technology Transfer Program.
NRC. 1990. Decline of the Sea Turtles: Causes and Prevention. Washington, DC: National Academy
Press.
Strauss, A. L., and J. Corbin. 1998. Basics of Qualitative Research: Techniques and Procedures for
Developing Grounded Theory. Thousand Oaks, CA: Sage.
Street, M. W. 1987. Sea Turtles and TEDs in North Carolina. Morehead City, NC: North Carolina
Division of Marine Fisheries.
Watson, J. 1980. Milestone Report: Sea Turtle Excluder Trawl Project. Pascagoula, MS: NMFS.
Weber, M. 1995. Chronology of Sea Turtles, Shrimp Fishing, and Turtle Excluder Devices. Washington,
DC: Center for Marine Conservation.
9 A Network States Approach for Mapping System Changes
Matthew M. Mehalik
Introduction
In this chapter, I describe the use of a three-states network framework, derived from
actor network theory, distributive cognitive systems, trading zones, and shared mental
models, to describe how a group of University of Pittsburgh researcher-interventionists
attempted to promote district alignment on several policy goals and measurement
initiatives in a large, urban school system in the United States. These researcher-
interventionists engaged various district personnel in a process of policy, measure-
ment, and network innovation.
I begin by describing the origins of the states framework in actor network theory,
shared mental models, distributed cognitive systems, trading zones, and boundary
objects. I then use the framework to describe how the researcher-interventionists
engaged with the school system. Finally, I interpret some of the findings from these
observations.
Because of their high degree of complexity, school districts are a perfect setting in
which to investigate the utility of such a framework. Two of the largest impediments
to district reform are the difficulty of aligning information systems and district data
with the decision-making capacities in the district, and the ways in which the district’s
extant internal sociotechnical networks become integrated with how the information
is collected, used, and implemented (Callon 1987; Latour 1999; Law 1987; Law and
Bijker 1992). The complexity of navigating these problems in network alignment is
also typically overshadowed by (but integrally linked with) the many challenges faced
by urban school districts. District personnel must make careful decisions about how
to direct their resources. Determining the correct policies to institute and knowing
how to move them forward are difficult precisely because of the network alignment
challenges.
182 Matthew M. Mehalik
The chapter describes how one district’s network changed over time, using a three-
states networking framework that has previously been applied to sociotechnical
engineering systems (Mehalik and Gorman 2006; Gorman and Mehalik 2002). The
framework, which will be discussed in detail later in the chapter, consists of three
network states:
• State 1, in which one actor or small elite group of actors has a comprehensive view
and “black-boxes” others into specific roles whose purpose(s) these others do not fully
understand;
• State 2, in which no group of actors has a comprehensive view. In this state, actors
construct and renegotiate “trading zones” (Galison 1997) and boundary objects (Star
and Griesemer 1989) that permit them to work together while still pursuing their own
enterprises or goals; and
• State 3, in which all participants share a common view (Gorman 1997).
Networks shift among these states. The framework is used as a tool to help practition-
ers reflect upon how to be strategic in their ability to work with others and to achieve
innovation amid network complexity.
The purpose for applying this framework to the urban school district case study is
to illustrate how such a framework can offer assistance to decision makers in complex
organizations who are responsible for initiating a process of innovation to improve
how that organization functions. In these cases, the decision maker must first be able
to understand what is currently happening in the organization, and then decide which
types of actions should be taken to achieve appropriate outcomes through the process
of change. In the parlance of trading zones, key questions for systems change include:
What sorts of trading zones exist in my complex organization? What holds them
together? For what reasons were they constructed in the first place? What new trading
zones should be fostered? How can they be created? The case study that follows is
intended to provide some insight into these questions.
Deriving the Three-States Framework
Actor Networks
Technologies do not … evolve under the impetus of some necessary inner technological or
scientific logic. They are not possessed of an inherent momentum. If they evolve or change, it
is because they have been pressed into shape. But the question then becomes: why did they
actually take the form that they did? (Bijker and Law 1992, 3)
An actor network perspective typically focuses on how human agents (superintendents

and principals, for example) try to strategically align other human agents (teachers
A Network States Approach for Mapping System Changes 183
and students) and nonhuman elements (such as technology and curricula) in a mutu-
ally sustaining way (Bijker, Hughes, and Pinch 1987; Bijker and Law 1992). A network’s
structure and alignment determine how well it responds to, deflects, or transforms
challenges to its integration (Law 1987, 132).
Latour and Woolgar (1979), Callon (1987), and Law (1987) have worked to create
a framework for discussing how new scientific facts emerge and how new technolo-
gies emerge out of rapidly changing networks of people and objects. In general, their
methodology consists of a “third way” descriptor system that attempts to avoid
debating whether social or natural effects are most important in determining how
new scientific facts and new technologies emerge. They have staked out this middle
ground to suggest a solution to the raging debate between most scientists—who
believe outsiders cannot study scientific knowledge because they do not understand
it, for they lack the training or insight to understand how nature plays a role in
deciding what is or is not a fact—and social relativists, who believe all knowledge is
of a social character and therefore can be subject to a sociological analysis in which
nature plays a lesser (or even no) role in arbitrating what should or should not be
called a scientific fact.
Callon, Latour, and Law use the concept of an actor network and a correspond-
ing methodology to escape this polarized and often acrimonious controversy. The
key to this methodology involves studying what is occurring at what Latour and
Woolgar (1979) call the “point of production.” At this point, a group of people and
objects attempt to configure themselves and one another in some way, according
to what Callon (1987) calls their “interests.” This conglomerate configures itself
according to what Latour (1999) calls an “association chain”—that is, some people
and objects become associated with one another in a type of linkage. The process
of this linking involves the recruitment of “actants” (the neutral term Latour and
Callon use to describe what they label as people and objects only after the point
of production). One key to understanding a recruitment strategy is paying attention
to how language and words are used, and also how actants juxtapose themselves in
relationship to one another. The next part of the point of production involves
whether the actants accept the way they have been juxtaposed—whether they remain
aligned according to the scheme of juxtaposition. If they do remain aligned in that
role, then they are considered recruited allies of the actor network. If they do not
remain aligned, they represent a potential contingency to the network that could
potentially influence other actants to withdraw from their roles—or disappear from
the actor network altogether. Callon (1987, 93–96) describes the process of accepting
roles or disappearing from the scene occurs as a “simplification process.” Among
the potentially infinite roles an actant could play, the actant remains in a single
role in relation to other actants. Latour and Woolgar (1979) refer to this simplifica-
tion process as “black-boxing.” Since those actants no longer represent a threat to
the network, their roles become taken for granted and are therefore black-boxed
(Bloor 1999).
The term “actor network” comes from the fact that this system of associations
is, as a whole, both an actor (forever changing, with the continual shuffling of
roles of actants) and a network (the actants aligned in a chain of associations). Law
(1987) refers to this process of production as “heterogeneous engineering,” that is,
the alignment of heterogeneous elements in an actor network. Those who study
actor networks must look at evidence at the point of production, because this evi-
dence permits the analyst to open any black boxes that have emerged from this
point of production. The analyst should also include himself or herself in the actor
network analysis, because the analyst acts as an actant in the network and therefore
affects the processes of juxtaposition and simplification of the actants in the actor
network.
The outcome of using this actor network framework is an insight into how facts
and technologies emerge from this process. The process of black-boxing strips away
all of the contingent associations among actants, so that the entity that emerges
appears to be free of contingency. For example, a statement such as “this school has
met its achievement goals” comes across as a fact when issued by state assessment
experts; however, such a statement hides the fact that there is a whole set of black-
boxed contingencies in the actor network. The statement “meeting achievement
goals” emerges from the alignment of an inspection by an expert who interprets
standardized test scores with an assessment protocol. But even these actants have
themselves been black-boxed into roles. What are the contingent factors associated
with creating the protocols? What are the contingent factors in deciding which levels
count as adequate progress? The opening of the black boxes reveals the contingencies
of earlier points of production that are hidden after the fact. Bijker and Law (1992)
call this stabilization “closure.”
Callon, Latour, and Law conclude that categories such as social forces and natural
forces are really just black boxes themselves that emerge after the point of production,
when an analyst looks back without attempting to open black boxes in her search for
the contingencies involving juxtaposition and simplification. Actor network theory
and methodology emerge as powerful tools for examining the process of how what
in the end become labeled “social and objective influences” affect how new ideas,
facts, and technologies emerge. Actor network theory is in other words a useful meth-
odology for understanding the dynamics of how actor networks change. Finally,
Callon (1989) and Bijker (1987) have used graphical representation techniques to
visually capture some of the network events previously described.
Some problems with actor network theory arise, however, because the role of an
individual’s agency often becomes lost in both the analysis and the individual’s per-
spective of the ability to be an active force in inducing change and innovation in
networks. If people are viewed as nothing more than “actants” in a large complex of
juxtapositions, and the chief role of such actants becomes a continual struggle to avoid
becoming black-boxed, it is very tempting to believe that individuals and groups
cannot take creative action to induce innovation and change in networks. And no
wonder innovation and change meant to help young people learn becomes such a
difficult prospect! Actor networks reduce the self-organizing and adapting nature of
people to too narrow a set of possibilities. Alignment as part of innovation is possible
without subjecting individuals to the constant threat of being black-boxed as actants.
Mental Models as Distributed Cognitive Systems

Gorman (1992, 1997) and Carlson (1991) attempted to follow the paths of three
inventors—Alexander Graham Bell, Thomas Edison, and Elisha Gray—of what later
became known as the telephone (Carlson and Gorman 1990). In undertaking this
study, they desired to gain insight into how creativity functions in the process of
invention while touching on the issue of determinism in invention. This problem,
debated often in both the social studies of science and technology and in the history
of technology, asks to what extent an invention is the result of the efforts of a creative
individual or the result of the larger social context in which that individual resides
(Bimber 1990; Staudenmaier 1985). In order to investigate this problem, Gorman and
Carlson chose a case of simultaneous invention, the invention of the telephone.
Interestingly, besides submitting patent documents for the telephone only a few hours
apart, Bell and Gray also submitted almost simultaneous patents for another invention
on a separate occasion. Explaining how simultaneous invention occurs is a main-
stream problem in the sociology of science and technology. To perhaps oversimplify,
most sociologists tend to point to contextual factors that took place in society at large
to explain why several individuals would come up with the same device at the same
time. The Strong Programme from Edinburgh makes the case that in both science and
technology, social interests dominate over any potential causal effects from nature
(Bijker, Hughes, and Pinch 1987; Bloor 1999). Another hypothesis is that Bell and
Gray stole ideas from one another, for patent agents communicated a potential inter-
ference in patent documents by pointing to a particular figure in a patent application,
which may have given Bell a tip on what to try next (Gorman et al. 1993).
Carlson and Gorman (1990) used a cognitive framework to examine a protocol of

notebook sketches, testimony, and correspondence to establish how ideas were com-
bined, reshuffled, discarded, synthesized, and embodied in actual devices. The frame-
work consisted of mental models (devices that could be run in the mind’s eye),
mechanical representations (stock solutions substituted into slots of the mental
model), and heuristics (rules of thumb). Bell, for example, based his speaking telegraph
on an analogy to nature, in this case the human ear, whose function he understood
in detail. Gorman et al. (1993) used a cognitive mapping scheme to trace changes in
these mental models over time. Inputs into the maps occurred from the influence of
outside social parties. The study’s framework showed that the inventors possessed
different representations for a telephone. Only in hindsight does it appear that they
invented the same thing.
Gorman (1997) followed up this study with the insight that Bell’s ideas were embed-
ded not only in his mind, but in his notebooks and in his benchtop devices as well.
That is, mental models are not just things that exist in the mind’s eye; they exist
embodied in the world of objects as well. Embodied knowledge could consist of tacit
knowledge, or craft skill (Collins 1985), as well as explicit knowledge such as that
written down, talked about, or drawn in notebook sketches.
The approach in Carlson and Gorman (1990) fits well with the studies reported by
Hutchins (1995). In his studies of ship navigation, Hutchins shows how both objects
and people can serve as means of distributing cognitive tasks associated with navigat-
ing a naval ship, such as computations and transformations, and information storage
(memory). Hutchins showed how even the simplest of devices, such as the markings
of a compass, are actually objects that have been constructed to aid in performing
cognitive functions and argued that they fit within a larger system in which those
computations and memory derive their meanings. In the example of the compass, the
markings are associated with a larger mental picture of a spherical world with a small
ship sailing upon it. These markings were made as a result of work by people in the
past, who discovered that a magnetic needle points to the north magnetic pole on a
spherical earth, and that this fact can be used to track a ship’s heading relative to this
mental picture. Yet the compass is one small part of the cognitive system of naviga-
tion. Charts, which are constructed as a bird’s-eye-view representation, are a conve-
nient way of relating the ship’s position to shorelines or depth levels. The bird’s-eye
view is not the actual position of the ship, but merely a representation that both stores
memory of where the ship currently is and suggests what should be done next in
terms of where the ship needs to go. The same results could be accomplished by
keeping a set of numeric coordinates for the ships and its obstacles, but numeric
coordinates present a heavy cognitive burden to anyone trying to process them. The
bird’s-eye-view navigational map reduces the cognitive load in the way it functions
as a device that embodies information. There is an observable correspondence between
what Hutchins (1995, 172–173) describes as distributive cognition and Gorman’s
(1997) concept of shared mental models and mental models embodied in objects.
Hutchins provides a way to link the roles of objects and social influences in a distrib-
uted cognitive system.
Figure 9.1 is a reconstruction of a diagram Hutchins (1995) uses to describe a snap-
shot of human practice at a given moment within a distributed cognitive system. The
horizontal axis represents the conduct of a given activity. Using the example of the
crew of a ship performing navigation tasks, a person manning the ship’s pelorus can
look at the shore, take a position reading, and then relay this information through a
communications system. This information is first spoken into the communications
system and is later translated into a written number, which is passed along to the
ship’s navigator, who takes the coordinate information and draws a mark on the ship’s
chart using plotting equipment. Along the “Conduct of Activity” axis, the operation
of the functional cognitive system brings representational media into coordination
with one another to perform that function. In this example, navigating the ship is
the function that is performed.
Simultaneous with the conduct of the activity, the practitioners themselves are
changing, or developing, their own representational media and capabilities. These
changes are often not explicitly articulated and exist in the form of changes in skills
and tacit knowledge. To use the ship navigation example, the crew members become
Development of Practitioners
Development of Practice
Conduct of Activity
Figure 9.1
A moment of human practice (redrawn from Hutchins 1995, 372).
better at using their equipment and make fewer mistakes with it. Such changes are
represented by the “Development of Practitioners” axis. Also occurring during the
activity are changes to the current organization of the social, material, and conceptual
aspects of the entire setting or frame of the cognitive system. This is shown on the
axis labeled “Development of Practice.” In Hutchins’s study of ship navigation,
changes to the navigation system occurred slowly and incrementally.
Hutchins also provides insight into how culture affects, and is affected by, these
simultaneous changes occurring within a moment of human practice and the social
and object embodiments of distributed cognition. He does this by contrasting the
way navigation is done by members of a Micronesian culture and the way naviga-
tion is done in Western culture aboard a U.S. Navy vessel. The Micronesians used
things such as the color of the water and the position of stars relative to observed
and imagined “reference” islands to help them navigate among the islands of Micro-
nesia. They had no need for a chart drawn from a bird’s-eye view, so there is no
corresponding mapmaker. They did not use compasses or a pelorus for determining
position, so neither these objects nor people trained to operate them exist in their
culture. Instead, there are people who are skilled at judging the color of the water,
looking for reference islands, and knowing the positions of various stars. In fact,
their entire representation of what they were doing is different from that in the
West, for the Micronesians used the cognitive construct that the ship remained still
and that the water, islands, and stars were moving relative to the ship. In the West,
the navigators take the earth and water to be remaining still, while the ship propels
itself through the water. Debating about who has the correct representation has no
meaning for navigation. Both forms of representations and associated distributed
systems of social roles and roles for objects are successful at meeting the goals of
the system.
Changes to the representative system along the “Development of Practice” axis
occur when the problem to be solved is discovered to be easier to solve using a differ-
ent type of representation, different object, or different social structure for processing.
In navigation, the chip log, which consisted of a rope with knots at even intervals
tied to a piece of wood, was replaced by the paddle wheel, which was replaced by
radar systems, which are in the process of being replaced by GPS satellite readings to
measure a ship’s velocity. This perspective is compatible with one of the conclusions
drawn from Kuhn (1996), who states that the purpose of scientific activity is to come
up with better ways of solving puzzles, often through reconceptualizing or reframing
the problem. Clearly the changes in the cognitive system over time take place to make
it possible to solve problems by using better representations.
Yet Kuhn’s hypothesis does not suit most of the systems studied by Hutchins,
because these are rather static in terms of change along the “Development of Practice”
axis. What about a design or invention situation where practice, practitioners, and
activity change in a complex dynamic interaction? Instead of looking at cognitive
systems that are coordinating their elements to perform a particular task, the Hutchins
framework can be tested in a situation involving rapid design and production of new
products using new design criteria for yet-to-be-developed markets. This is exactly the
situation in the case of interventions in school systems. To date, this problem of rapid
changes in social and cognitive systems has not been adequately addressed by any of
the scholars mentioned so far in this chapter, although the tools generated by Callon,
Latour, Law, Hutchins, Gorman, and Carlson may be useful for solving this task.
Therefore, a cognitive system analysis that combines the Hutchins framework for
capturing the operation of functional cognitive systems with the concepts of embod-
ied mental models as discussed by Gorman (1997), and the network alignment,
juxtaposition, black-boxing, and closure processes described by Latour and Woolgar
(1979), Callon (1987), and Law (1987) can be useful in addressing how interventionist
innovation in school systems can proceed.
Boundary Objects and Trading Zones
Two additional useful concepts for describing how systems of objects and people
organize themselves in performing functional tasks are boundary objects (Star and
Griesemer 1989) and trading zones. Galison (1997) describes the role of objects as
serving as a type of glue to bind together cultural practice in the field of experimental
physics. In Galison’s description, objects serve as a way to preserve communication
capabilities among various actors who are trying to communicate across distinct
knowledge domains. In terms of shared mental models and Hutchins’s idea of evolu-
tion along the “Development of Practice” axis, Galison’s objects permit sharing among
people who do not necessarily hold the same mental model or who operate using
different conceptions of practice. For purposes of the framework I introduce in the
next section, I use the term “boundary object” in the sense that Galison implies.
Galison (1997) also uses the concept of a trading zone to discuss cultural behavior
in which certain individuals are capable of communicating with other individuals
in a separate knowledge or functional domain. These trading zones are constructed
by negotiating the limits of what constitutes a valid trade and the limits and bound-
aries of knowledge transfer. Collins, Evans, and Gorman (this volume) provide a
framework that shows how expertise serves as a key dimension of how trading zones
are organized. Their focus on expertise provides a grounded and practical approach
for describing the creation of network alignments and trading zones.
Network States Framework
We now have before us the goal of finding a way to capture and discuss some of the
characteristics of dynamic knowledge and computational functional systems by inte-
grating some of the concepts from the discussion in the last section. Many of these
concepts captured particular aspects of various network and cognitive system pro-
cesses. The integration of the concepts into a common framework may help to meet
the overall goal of this chapter, which is to discuss how urban school systems are able
to construct tools, products, processes, and network relationships in pursuit of various
improvements in learning policies and practices.
Clearly, networks align themselves for many reasons. Actor network theory accounts
for many of these processes: alignments according to social power, resources, network
loyalty, among others. Distributed cognitive systems illustrate how knowledge becomes
aligned. Collins, Evans, and Gorman illustrate how expertise plays a role in these
alignments. The goal of this chapter, however, is to offer a framework and tool that
can be useful to decision makers who are in the midst of trying to manage complex-
ity—situations in which a thorough and comprehensive inventory of knowledge flows,
expertise, and complete social dynamics is not possible. For this reason, my emphasis
is on utility rather than comprehensiveness.
As discussed earlier, this integrated framework (Mehalik and Gorman 2001; Gorman
and Mehalik 2002; Mehalik and Gorman 2006) consists of three network states (state
1, state 2, state 3). In a state 1 network, one actor or a small elite group of actors has
an overall representation of the problem, or mental model. This one actor black-boxes
others into specific roles whose purpose those persons only partly understand, and
sets parameters of any trading zones among the actors. In practice, state 1 networks
function in a top-down configuration. An example of a state 1 network is Pizarro’s
conquest of the Inca Empire, in which he tricked the Inca Emperor Atahualpa, killed
him, and then imposed his rule upon the Inca civilization (Diamond 1997).
In a state 2 network, no group of actors has a comprehensive view or overall rep-
resentation of the problem. Each actor is connected to others by a boundary object,
which each actor sees or understands differently. The individual actors can be pursu-
ing their own enterprises in relation to a common boundary object. Successful state
2 networks include active trading zones, in which the boundaries of what constitutes
an allowable “trade” are negotiated among the actors. Unsuccessful networks include
ones in which the language or boundary object connecting the trading zone fails, or
the boundaries of the trading zone are violated by one of the actors. This situation
forces the network to shift to a different state or to disassociate altogether. Most rela-
tionships tend to exist in a form of a state 2 network. For example, a supplier-customer
relationship is governed by the existence of negotiated limits on what constitutes an
active trade (say, a supplied product in exchange for payment). If the limits of this
trade are violated, the relationship may dissolve.
Finally, in state 3 networks, all participants share a common representation or
mental model. Because there is open sharing among all members of a state 3 network,
trading zone boundaries are not openly negotiated. In a dynamic network situation,
there needs to be continuous sharing so that a common representation is maintained.
This concept is similar to the idea of a shared mental model. An example of a state 3
network is the inner circle of practitioners who constructed the ARPANET (Hughes
1998).
This three-states framework maps directly onto the Hutchins (1995) framework
depicting a moment of practice (figure 9.2). The framework provides a way of talking
about rapidly evolving functional cognitive systems in addition to stable ones, such
as ship navigation.
The three-states framework also maps onto a visual framework that can be used to
describe changes in the network. Figure 9.3 shows two actors, designated by ovals,
who are interacting in a network. The porous circular line surrounding an actor,
or group of actors and objects, indicates that they are all organized in a state 3
Development of Practitioners
— Changes in motives, goals,
skills, tools through state 2
processes
Development of Practice
— Changes to shared
mental models
— State 1 or 3 shifts
via state 2
Conduct of Activity
— Stable or slowly evolving
state 3, 2, or 1
cognitive system
Figure 9.2
States framework mapped onto Hutchins’s (1995) moment of practice.
State 2 trading zone

in operation
(with boundary object)
Mental
Model
Boundary State 2 trading zone
Object just opening
Actor
through negotiation
(no boundary object)
State 3 boundary
(stable mental
model/distributed Actor
cognitive system
boundary) Unidirectional Mental
trading zone Model
(state 1)
Figure 9.3
Cognitive mapping convention.
configuration. This means that all items within that circle are organized according to
a shared mental model, with no trading zones necessary within that state 3 network.
The porous boundary indicates that the network is a slowly evolving cognitive system
in terms of the development of practice. The navigation systems that Hutchins studied
constitute such a situation. Within the state 3 network, it is possible to examine the
processing and computations that the network is performing to accomplish its well-
specified functional objective.
The lines running between the circular boundaries indicate trading zones between
different actors and networks. Arrows pointing in only one direction indicate that the
trade takes place according to trading zone boundary rules that are determined by one
actor or network and are being imposed on another network. In other words, unidi-
rectional arrows indicate a state 1 relationship between the actors sharing the unidi-
rectional arrow.
State 2 trading zones are shown using bidirectional arrows. When a trading zone
first opens, the boundaries are still being negotiated. This type of a situation is shown
using a dotted bidirectional arrow. When the state 2 trading zone has stabilized
through the use of a boundary object, which cements trading zone boundaries, the
bidirectional arrows point to the boundary object, and the arrows are shown in solid
print. This combination of the three-states framework and the cognitive mapping
conventions permits an analysis of a dynamic network. The framework will also

provide some structure for my discussion of how a network using a systems-mapping
methodology has been able to engage in policy innovation in a large, urban school
system in the United States.
Use of the Framework to Interpret an Intervention in an Urban School District
Alignment is a central leadership challenge for transforming systems, especially soft

systems such as school districts. By “alignment,” I mean the ability of the system to
function as a structured and dynamic whole, with all elements of the system sharing
a relation to a common identity and/or purpose (Callon 1987). For education systems,
the challenge involves aligning—or, in some cases, realigning—all aspects of the
system to the common purpose of increasing the learning capacity for young people
in key knowledge domains, such as math, reading, writing, and science. This challenge
is difficult for many reasons, one of which is that it is difficult for different people
with different areas of expertise, and working in different areas of an organization, to
relate to any new idea, object, person, or initiative in a way that typically creates an
alignment within the network.
This central leadership challenge can be broken down into the need to provide the
necessary structures of intervention for both the more directed and constrained por-
tions of the system and the more autonomous and self-organizing aspects of the
system. (All complex systems have both elements.) There are several aspects of this
leadership challenge: (1) figuring out which aspects of the system function more
effectively and efficiently when directed/constrained and which function better when
they are more autonomous; (2) discovering and removing bottlenecks, unnecessary
redundancies, and resource allocation for the constrained portion of the system; (3)
providing relevant information, policy affordances, and opportunities for developing
shared meaning for the self-organizing portion of the system; and (4) embedding the
most effective accountability structures in both dimensions of the system so that
alignment and appropriate adaptation can be examined, studied, maintained, or
redirected.
The constraints and enablers that tend to flow from higher levels in district systems
correspond to the extent to which policies and work practices either limit or encour-
age what people within each of the system’s levels do. Within each of the levels, actors
are able to continue to follow existing scripts, can modify their scripts when given
changes to what is happening within the level in which they are working, and/or can
discard existing roles and routines in order to meet the local, embedded circumstances
that actors at those levels perceive as necessary in order to meet their goals. Coburn
(2005) has demonstrated that this interplay of the directed and autonomous aspects
of school systems typically results in the failure to embed the intended goal of a policy
into practice. In other words, the directed aspect of the district exerts only so much
influence on the autonomous, local practice. This occurs for several reasons, including
lack of realistic anticipation and design in the process by which the new policies are
created, lack of affordances in the handoff of policies to practitioners at various levels,
lack of knowledge on the part of local actors who encounter the new policy, lack of
interest or difficulties in changing routines, and perceived threats to identity and
autonomy on the part of actors with respect to the policy’s intentions or goals.
Spillane (2006) argues that it is necessary to look beyond the factors of individual
agency and the influence of social structure, and instead focus on interactions and
practices, such as organizational routines, using useful frameworks for practitioners.
It is possible to accomplish this call to action while paying attention to how such
interactions become aligned with network features.
Figure 9.4 shows the organizational chart of the large, urban school district in
which the University of Pittsburgh researcher-interventionists began their work of
using systems mapping. The initial contact between the district and the research team
began with a mutual interest in finding ways to measure student achievement for the
district as a whole. The district administration was in the process of constructing a
“scorecard” of items believed to be worth measuring in the district. The measuring of
those factors would produce useful data that could inform the district’s ability to make
informed decisions. In this initial period, the project resided with the director of
assessment (DA) and the director of technology services (DTS). The DA was bringing
to the project knowledge about assessment strategies and techniques that were in
operation in the district as well as ideas about which of these might be worth assessing
for the scorecard. The DTS was familiar with the extensive array of databases, knew
how information was and could be stored, retrieved, and analyzed, and knew the
quality of data that the district was already collecting.
Initially, the University of Pittsburgh team interacted with the district only through
a postdoctoral researcher (not the author) who in 2004 was helping to determine what
might be worth measuring from the perspective of promoting equitable student
achievement in the district. It became apparent that trying to focus on only one
dimension (equity) amid the complex nature of school systems and the factors that
contributed to student achievement required looking at the district from a larger
perspective. It also required figuring out which factors in school systems influenced
both student achievement outcomes and equity in achievement. The Pitt team
recruited more members—three industrial engineers, a highly experienced education
Chief of Staff
Board of Education
Special Asst. to
Superintendent
Foundation
Superintendent
President
Communications
Academic Policy
Advisor
CFO COO CAO General Counsel
Div. Teaching
Dir. Budgeting Spec. Asst. to and Learning
COO
Dir. Early Ed. Dir. Humanities Instructional Exec. Dir.
Risk Superintendents Student Svcs.
Management Dir. Human
Resources Dir. Extended Dir. Div. Instructional
Learning Math/Science Support Dir. Special
Controller Education
Technology
Services Dir. Interdisc. Dir. Eng Lang
Curriculum Acq Dir. Prevention
and Intervention
A Network States Approach for Mapping System Changes
Purchasing
Dir. Planning Dir. Leadership Div. Student
and Research and Development Services
Facility
Management
Dir. Dir. Career and
Assessment Tech. Ed.
Payroll
Dir. Alt Ed. Dir. Library

Security and Media
Dir. School
Transportation Choice Dir. Accountability
Figure 9.4
District organization chart.
195
researcher who knew a lot about research on impacts on student achievement, and
an additional education researcher who focused on leadership in school systems.
In the summer of 2005 the Pitt team spent several months interviewing experts
and reviewing literature in an attempt to determine which factors were most proximal
to having an effect on student achievement. The team also identified different factors
at the classroom, school, district, and state levels, and arranged these factors in causal
chains based on expert interviews. From this process, the Pitt team created a “theoreti-
cal system causal map.”
Part of conducting any system analysis involves using tools to help engage clients
in discussions to elicit useful information about what the client knows about a par-
ticular topic, so that the analyst can design a solution that suits the client’s needs.
For this purpose, the Pitt team decided to transform their map into an exercise that
could help school district personnel develop their own hypotheses about how factors
in their district related to one another causally and impacted student learning. The
early versions of the maps and the exercise were shared with various school district
personnel at a retreat to solicit their expertise on the structure and utility of the maps
and the mapping exercise. A prototype of the activity was then piloted in a medium-
sized urban district in the United States in the winter of 2005. In the spring of 2006,
the Pitt team also introduced this activity to the DA and DTS in the large, urban
school district working on the scorecard assessment project. Both the DTS and DA
initially resisted the Pitt version of the causal maps and instead insisted on deriving
their own version of the map, which the Pitt personnel then reconciled with their
version.
However, it then became unclear what benefit this mapping activity would be able
to deliver to the district. The DTS was tasked with completing the measures of the
scorecard initiative, and that task remained the highest priority. Although the Pitt
team was interested in assisting in the scorecard’s construction, the team also wanted
to use the mapping tool as a way to describe the complexity within school systems
and as a useful tool for leaders who wanted to improve student performance in school
systems. Figure 9.5 illustrates the situation that unfolded: a state 2 trading zone was
in place, in which no group of actors had a comprehensive view, and the scorecard
effort remained the boundary object linking the school district and the Pitt team.
From the Pitt perspective, this settled state 2 trading zone included the swapping of
some skills, expertise, and insight from the causal system maps in exchange for the
ability to begin modeling some of the causal chains using district data. The district
benefited from having access to experts who could assist in clarifying some of the
measures and linkages in the scorecard assessment activity.
Chief of Staff
Board of Education
Special Asst. to
Superintendent
Superintendent
Communications
Academic Policy
Advisor
Div. Teaching
COO
Dir. Early Ed. Instructional Exec. Dir.
Risk Superintendents Student Svcs.
Resources Dir. Extended
Learning
Controller
Technology
Services Dir. Interdisc.
Curriculum
• System Measures
• System Measures Dir. Planning • Equity Measures
• District Plan and Research
• Mapping System Structure
• Useful
• Equity Measures Dir.
Assessment Pitt Systems
Modeling
Group
Scorecard
Assessment
Figure 9.5
Network stabilizes around scorecard initiative.
197
The network changed when one of the Pitt team members met with the district’s
academic policy advisor (APA). The APA was tasked with trying to ramp up several
policies for the superintendent, who had joined the district in the previous calendar
year. These policies included finding ways to improve math and reading achievements
with new curricular designs, the boosting of student engagement, and the introduc-
tion of new professional development that included the use of academic coaches in
core subject areas. The APA asked the Pitt team if they could help design these policies
in such a way that their effectiveness could be measured, providing rapid feedback
from the initial versions of the policies as they were introduced to key management
personnel throughout the district. In other words, the APA was thinking about the
ability to align the district to these new policies.
The Pitt team offered the use of the mapping activities as a way to engage these
key management personnel in suggesting useful ways of measuring whether these new
policies were achieving what they were intended to achieve. By the last quarter of
2006, the Pitt team settled into a stable state 2 network in which there were now two
activities functioning as boundary objects: the scorecard initiative and the policy
construction initiative. The Pitt team was gaining access to more of the district’s
information and personnel in order to effect change using the mapping activities and
expertise in data acquisition, modeling, and network assessment. The district was
gaining access to expertise that would involve the construction of meaningful mea-
sures for the scorecard initiative and for the new policies. From the perspective of
alignment of the network, the APA, DTS, and the Pitt team were all working toward
the same goals of obtaining useful measures for the district in terms of district policies
and academic achievement. Each of these groups also had the common goal of helping
other parts of the district’s network become aligned to the task of generating useful
measures for new policies and for the scorecard.
However, the APA, DTS, and the Pitt team were approaching this issue of alignment
from separate shared mental models. The Pitt team was interested in validating the
mapping activity and their ability to test advanced data-modeling techniques in the
district. Thus, their mental model was weighted toward research and intervention.
The APA’s mental model was related to the district superintendent’s goals for launch-
ing the new policies in academic achievement, student engagement, and coaching/
professional development. Finally, the DTS operated from the mental model of getting
the scorecard initiative operational with effective data measures. The boundary objects
of the policy construction task and the scorecard initiative task permitted the groups
to align their efforts without requiring the mental models of each of the groups to
shift. Figure 9.6 shows the network in this settled state 2 configuration, with the two
boundary objects linking the different members of the network together.
Chief of Staff
Board of Education
Special Asst. to
Superintendent • Student Achievement
• Student Engagement
Superintendent
• Professional Devt.
Communications
• Others
• Quick Turnaround
Academic Policy • Data Measures
Advisor
Div. Teaching
COO
Dir. Early Ed. Policy
Risk Construction
Resources Dir. Extended
Learning
Controller
Technology
Services Dir. Interdisc.
Curriculum
• System Measures
• System Measures Dir. Planning • Equity Measures
• District Plan and Research
• Useful
Modeling
Group
Scorecard
Assessment
Figure 9.6
199
Network encounters policy construction challenge.

In late 2006, the Pitt team held several meetings with the district’s director of
leadership and development (DLD), the director of math and science (DMS), and the
director of humanities (DH). In each of these meetings, the Pitt team mentioned the
possibility of using the mapping task to help each of these leaders clarify the causal
linkages involved with the release of new curricular guides and for the introduction
of coaches to support the new curricular initiatives. None of these three leaders chose
to engage in the mapping activity, although all three understood the need to help
achieve alignment in the district on the introduction of the new curricula. Yet,
although the mapping activity itself failed to act as a boundary object for these per-
sonnel, the policy initiatives and scorecard initiative did serve as boundary objects
between these leaders and the Pitt team. The “trade” in this case consisted of the Pitt
team’s agreement to provide insight into how the network of coaches, instructional
superintendents, and instructional support staff was managing to introduce the new
curricular guides to the district, as well as to suggest possible ways of measuring their
adoption and implementation, so that some sense of uniformity could be achieved.
The Pitt team thereby was granted access to additional parts of the network, such as
the opportunity to meet with school district principals and data assessment teams in
professional development sessions. The results of these meetings, which took place in
the winter of 2006 and spring of 2007, were too recent to be included in the present
analysis. Figure 9.7 illustrates the interactions of these personnel, with the policy
initiatives and the scorecard initiatives serving as boundary objects.
It was not until early 2007 that the Pitt team met with the district superintendent
and the chief academic officer (CAO) to explain in person the purpose of their activi-
ties. Until that point, neither the superintendent nor the CAO had known the extent
of the Pitt team’s involvement with the district, particularly their work on developing
measures for the scorecard initiative. This situation did not present a problem, mostly
because the APA had ensured that the Pitt team’s efforts were in direct alignment with
the superintendent’s and the CAO’s goals for the new policy introductions and the
ongoing efforts to complete the scorecard, both priority items of the superintendent.
It became clear at that time that the superintendent, the CAO, and the APA were
operating with a common mental model for how changes in the district were to
proceed with respect to the scorecard and the policy changes in achievement, engage-
ment, and professional development/coaching. That mental model could be described
as the desire to quickly implement measurable policies and procedures that would
produce useful information to aid the superintendent and the CAO in their decision
making. From the perspective of the Pitt team, it became clear that the superintendent,
the CAO, and the APA operated in a state 3 network regarding these policy issues; that
Chief of Staff
Board of Education
Special Asst. to
Superintendent • Student Achievement
Superintendent
• Professional Devt.
Communications
• Others
• Quick Turnaround
Academic Policy • Data Measures
Advisor
Div. Teaching
COO
Dir. Early Ed. Dir. Humanities Policy
Risk Construction
Resources Dir. Extended Dir.
Learning Math/Science
Controller
Technology
Curriculum Acq
• System Measures
• System Measures Dir. Planning Dir. Leadership • Equity Measures
• District Plan and Research and Development
• Useful
Modeling
Group
Maps and
Scorecard
Assessment
Figure 9.7
201
Mapping policies for district alignment.

is, all participants shared a common mental model. This situation is shown in
figure 9.8 at the top of the figure.
The Pitt team also began to convene meetings with several of the special assistants
to the superintendent (SATSs) in the areas of student engagement and academic
achievement. Again, the policy initiatives and the need to create measurable items for
the scorecard effort served as boundary objects for the interactions with the SATSs.
With these interactions, however, there were follow-up meetings that included staff
members under the direction of each SATS. At these meetings, the Pitt team was able
to engage these staff members in analyses of the causal factors that impacted student
achievement from the perspective of different enablers. In one case, the focus was on
the causal factors that impacted student engagement. In another case, the focus was
on the causal factors that impacted community involvement as part of student engage-
ment. Another meeting focused on academic achievement from the perspective of
instruction. In all three cases, the groups came together to construct maps of how
different school system factors impacted their areas of focus. Each group generated
what they believed to be measurable indicators that would be useful for determining
how these factors could be changed in order to achieve desired outcomes. These
mapping exercises resulted in information that was useful for both the scorecard ini-
tiative and the policy initiative. Furthermore, the mapping task served as a way to
elicit input from district personnel at different levels within the management portion
of the district. Thus, it served as a tool to encourage each group to generate and share
mental models for how the district worked.
The maps therefore were a useful tool for inducing the groups into a state 3
network for the short time they were engaged in the mapping activity. As such, the
maps fulfilled the role that Hutchins (1995) called the “residua of the process”—in
other words, they were an artifact of the activity of creating a shared representation
(and perhaps shared meaning) of how causal factors in the district acted upon the
participants. The maps also served as long-term memory devices, and perhaps as
boundary objects, once the group members returned to their separate places within
the district when the activity concluded (I have not had the opportunity to observe
evidence that the maps did in fact fulfill such a capacity directly). The meetings
themselves permitted the participants to propose new ways of working together
(creating new trading zones) within the groups. It will be interesting to follow how
the dynamics of these groups change over time. At the moment, the activity has
been too recent for such results to be observed. However, we can see that the maps
did serve as a way to achieve alignment in the district with respect to the need to
construct measurable indicators for the scorecard initiative and for the new policy
Chief of Staff
Board of Education
General Counsel
• Student Achievement
Superintendent
Communications • Professional Devt.
• Others
Academic Policy • Quick Turnaround
Advisor • Data Measures
Special Asst. to
CFO COO CAO
Superintendent
Div. Teaching
COO
Dir. Early Ed. Dir. Humanities Policy
Risk Construction
Resources Dir. Extended Dir.
Learning Math/Science
Controller
Technology
Curriculum Acq
• System Measures
• System Measures Dir. Planning Dir. Leadership • Equity Measures
• District Plan and Research and Development
• Useful
Modeling
Group
Maps and
Scorecard
Assessment
Figure 9.8
203
Policies for district alignment (including leadership).

initiatives. They did serve as a tool to induce local networks into a state 3 situation,
in which the members of the network could construct meaningful, common measures
relating to the policy initiatives. The mapping task did help these networks to con-
tinue to revisit what they had created so that they could improve and revise their
measures and representations. Finally, the mapping task permitted the more local
aspects of a network to contribute their perspectives and insights to a process that
in fact aligned with the vision and initiatives generated by the higher management
of the district.
A Different View of District Alignment
The mapping techniques described in this chapter have highlighted how a group of
practitioners from multiple areas in a school system can articulate the flow of interac-
tions that eventually lead to classroom activities that impact student learning. Once
practitioners described their views of how different aspects of the system worked, their
conversations were directed to determining what factors in the system were the most
constraining. Just as practitioners in production systems can usually point to places
that are bottlenecks in those systems, participants in the mapping exercises created
hypotheses for examining where bottlenecks were in their district, and what was worth
measuring in their district. Once such items became visible, it became possible to
suggest how the resources, roles, and routines of people within the system could be
configured so that bottlenecks could be removed or their capacity increased.
The participatory nature of the mapping exercise also gave practitioners the ability
to construct a shared meaning of how their district works, and also how their roles
could change with respect to making their district work more effectively. The common
perspective of aligning toward meeting the primary goal of the system promotes vested
interest in seeing that system improve.
Accountability in such a district becomes more than simply a metric of enforcement
and potential punishment. Accountability refers in this case to the adaptive mecha-
nisms put into place that permit actors within education systems to make changes to
resources, routines, and practices that promote the performance of the system in
which all actors have a stake. Accountability can include the participation of teachers
and staff as members of data teams that focus on academic subject areas. The teams
can meet to discuss the implementation of new curricula and strategies to meet aca-
demic objectives. They can also discuss embedded, formative assessments that measure
classroom learning on topics. As part of their discussions, such teams can determine
the constraining factors of these formative assessments, and these needs can be
communicated to actors in the system so that resources can be deployed to remove

such bottlenecks in the system.
Data teams also involve the coming together of people in order to find common
meaning in their practices so that they can discover how their practice can evolve to
become more effective. Common meaning is created around the measures of assess-
ment; as a result, accountability reflects the interests of the people who are being held
accountable, because the accountability measures reflect the commonly aligned goals
for the district. The accountability structures will also mirror the dimensions of dis-
tricts that operate as constrained/directed, as well as the more autonomous aspects.
Both dimensions of districts will have measures of accountability; however, the par-
ticipatory practice of creating many such measures can hopefully promote the realiza-
tion that, whether they speak to the constrained aspect of the district, or whether they
are tied to local, more autonomous practices, all of the measures speak to alignment
to improve student academic performance, and they all provide insight as to where
key bottlenecks exist and how resources can be redeployed to remove them.
The mapping and data collection can also make the actors’ roles in the district
clearer—and can make their lines of work more meaningful by eliminating much of
the second-guessing, the sense of powerlessness, and the resulting perception of the
need to construct miniature, semiautonomous “kingdoms” that waste resources; in
their place, the mapping and data collection provide mechanisms for addressing the
need for change. The flip side of this process is that data collection and measurement
become embedded in meaningful lines of work. Data collection for its own sake does
not guarantee improvements in system performance—in this case, student academic
learning. Indeed, hospital systems that measure process performance without embed-
ding the data collection into meaningful work practices have shown that such data
collection efforts do not tend to be associated with improvement in the hospital
systems (Werner and Bradlow 2006). When used in strategically participative ways,
system tools, such as mapping process flows, can promote recognition by district actors
of their role within the larger system and can help actors recognize how resources
affect their ability to be effective by being directed/constrained in certain respects and
autonomous in other respects. And when they experience these two dimensions of
their practice, they realize that the differences can be justified by the need to allocate
resources in ways that most effectively promote student learning outcomes, a process
that the actors are involved in constructing and that requires measurement.
This chapter has used a states framework to illustrate how alignment can be
achieved in a school district, even though the entire district at a given moment may
not share exactly the same meaning that the superintendent directs for a district. The
case discussed in this chapter shows that it is possible to promote alignment within
a district by inducing shifts into state 3 and state 2 network configurations as is stra-
tegically advantageous. (The chapter did not include examples of state 1 aspects of
the network simply because no such states provided insight into the interactions
described in this chapter. This does not mean that such network processes were absent
in the school district or even within the Pitt team itself.) Instead of lamenting that
the school policies that get implemented are seldom the school policies that were
designed, the researchers focus on what tools can prove useful to promote network
shifts into different states that will accomplish the overall goals of a school system (or
other complex sociotechnical network). The mapping exercises and tools described in
this chapter are a stepping-off point for what will be useful tools to promote alignment
with higher student achievement in school systems.
From the perspective of trading zones, this chapter showed how intervention in a
complex system did not require a comprehensive mapping in great depth of trading
zones. For decision makers and persons responsible for achieving effective systems
change, such a detailed inventory was not possible due to time and resource con-
straints. Typically, because of these time and resource constraints, administrators tend
to rely on power configurations in order to command changes. Such configurations
tend to narrow trading zones across districts; however, enforcement of this narrowing
becomes difficult. What happens in these instances is the well-documented phenom-
enon that “the policy commanded is not the policy enacted” (Coburn 2005; Spillane
2006).
Instead, this case illustrates that it is possible to use a tool in the form of the three-
states network to intervene strategically in complex systems in order to induce systems
change through the construction of different trading zones. The states framework
provided a rough glimpse into the trading zones that held together different entities
in a network. These trading zones can be based upon concepts of expertise as well as
power, loyalty, resources, and status, among other factors. In the heat of a decision-
making moment, it may be unnecessary to go into this degree of detail. What is
important is the ability to document what is held together. What may be important
is what the next move will be to change the network. Change may not require the
construction of additional trading zones. In fact, this case illustrates the power of
bringing together groups into a state 3 condition, in which various participants create
meaning together through a sequence of mapping exercises. Once this process was
completed, each of these network members was able to go off and renegotiate their
own sets of trading zones in state 2 network conditions in other parts of the network.
The construction of meaning in a state 3 network is an important dynamic that the
trading zone, as a concept, tends not to capture. And the construction of trading zones
in state 2 networks can occur due to a whole host of factors (expertise, power, loyalty,
resources, status, among others). The three-states framework is intended to provide a
tool that is useful and does not carry too much of a cognitive burden while being
open to more detailed analyses.
Indeed, follow-up research should study real-time, benchtop, and negotiation pro-
cesses that are necessary to capture fine-grained, moment-by-moment cognitive con-
structions in multiple environments. Such research can involve developing indicators
and observation protocols of the parameters of trading zones, such as expertise, power,
and status. Such a study can eventually result in a simulation for training decision
makers who find themselves responsible for managing complex and changing socio-
technical systems.
References
Bijker, Wiebe E. 1987. The Social Construction of Bakelite: Toward a Theory of Invention. In
Bijker, Hughes, and Pinch 1987.
Bijker, Wiebe E., Thomas P. Hughes, and Trevor Pinch. 1987. The Social Construction of Technologi-
cal Systems: New Directions in the Sociology and History of Technology. Cambridge, MA: MIT Press.
Bijker, Wiebe E., and John Law, eds. 1992. Shaping Technology/Building Society: Studies in Sociotech-
nical Change. Cambridge, MA: MIT Press.
Bimber, Bruce. 1990. Karl Marx and the Three Faces of Technological Determinism. Social Studies
of Science 20:333–351.
Bloor, David. 1999. Anti-Latour. Studies in History and Philosophy of Science 30 (1):81–112.
Callon, Michel. 1987. Society in the Making: The Study of Technology as a Tool for Sociological
Analysis. In Bijker, Hughes, and Pinch 1987, 83–103. Cambridge, MA: MIT Press.
Carlson, W. Bernard. 1991. Innovation as a Social Process. Cambridge: Cambridge University

Press.
Carlson, W. Bernard, and Michael E. Gorman. 1990. Understanding Invention as a Cognitive

Process: The Case of Thomas Edison and Early Motion Pictures, 1888–1891. Social Studies of Science
20 (3):387–430.
Collins, H. M. 1985. Replicating the TEA-Laser. In Collins, Changing Order: Replication and Induc-
tion in Scientific Practice, 51–78. Beverley Hills and London: Sage. [2nd ed., Chicago: University
of Chicago Press, 1992.]
Coburn, Cynthia E. 2005. Shaping Teacher Sensemaking: School Leaders and the Enactment of
Reading Policy. Educational Policy 19 (3):476–509.
Diamond, J. 1997. Guns, Germs, and Steel. New York: W. W. Norton.

Galison, P. 1997. Image and Logic. Chicago: University of Chicago Press.
Gorman, Michael E. 1992. Simulating Science. Bloomington: Indiana University Press.
Gorman, Michael E. 1997. Mind in the World: Cognition and Practice in the Invention of the
Telephone. Social Studies of Science 27:583–624.
Gorman, Michael E., and Matthew Mehalik. 2002. Turning Good into Gold: A Comparative Study
of Two Environmental Invention Networks. Science, Technology and Human Values 27 (4):499–529.
Gorman, M. E., M. M. Mehalik, W. B. Carlson, and M. Oblon. 1993, Alexander Graham Bell,
Elisha Gray and the Speaking Telegraph: A Cognitive Comparison. History of Technology
15:1–56.
Hughes, Thomas P. 1998. Rescuing Prometheus. New York: Pantheon.
Hutchins, Edwin. 1995. Cognition in the Wild. Cambridge, MA: MIT Press.
Kuhn, Thomas S. 1996. The Structure of Scientific Revolutions. 3rd ed. Chicago: University of
Chicago Press.
Latour, Bruno. 1999. For David Bloor … and Beyond: A Reply to David Bloor’s “Anti-Latour.”
Studies in History and Philosophy of Science 30 (1):113–129.
Latour, Bruno, and Steve Woolgar. 1979. Laboratory Life: The Social Construction of Scientific Facts.
Beverly Hills: Sage Publications.
Law, John. 1987. Technology and Heterogeneous Engineering: The Case of Portuguese Expan-
sion. In Bijker, Hughes, and Pinch 1987.
Law, J., and W. E. Bijker. 1992. Postscript: Technology, Stability and Social Theory. In Bijker and
Law 1992.
Mehalik, M. M., and M. E. Gorman. 2001. Understanding Moral Imagination: Applying a Network
State Framework to Cases of Inventing for the Environment. Proceedings of the 2001 American
Society for Engineering Education Annual Conference and Exposition, Albuquerque, New Mexico.
Mehalik, M. M., and M. E. Gorman. 2006. A Framework for Strategic Network Design Assessment,
Decision Making, and Moral Imagination. “Engineering Ethics.” Special issue, Science, Technology
and Human Values. May.
Spillane, J. 2006. Distributed Leadership. San Francisco: Jossey-Bass.
Staudenmaier, John. 1985. Technology’s Storytellers. Cambridge, MA: MIT Press.
Werner, Rachel M., and Eric T. Bradlow. 2006. Relationship between Medicare’s Hospital Compare
Performance Measures and Mortality Rates. Journal of the American Medical Association 296
(22):2694–2702.
10 Embedding the Humanities in Engineering: Art, Dialogue, and a
Laboratory
Erik Fisher and Roop L. Mahajan
Introduction
In this chapter, we discuss the development and pursuit of two interdisciplinary

trading zones in which the authors participated: (1) an initial year in which we
developed the notion of “humanistic engineering” in the University of Colorado at
Boulder’s College of Engineering and Applied Science, and (2) a thirty-three-month
period in which Fisher functioned as an “embedded humanist” in Mahajan’s Thermal
and Nanotechnology Laboratory. In both cases, we sought to integrate the divergent
perspectives of engineering and the humanities in order to enhance the ability of
engineers—in undergraduate, graduate, and ultimately professional contexts—to
engage in productive, self-critical inquiry.
After describing areas of overlap between our various backgrounds and perspec-
tives, we describe our collaborative undertakings, including the genesis and employ-
ment of several locutions and metaphors that framed and facilitated our efforts. We
reflect on these collaborations in relation to a combined framework of trading zones
and interactional expertise (Collins, Evans, and Gorman 2007). We outline the lin-
guistic terms and metaphors we employed during our endeavors in order to recon-
struct the coevolution of our thinking during the development and structuring of
our interactions. Further, we consider the role played by cognitive activity in our
account of these interdisciplinary collaborations. As will be evident, our collabora-
tions were at times propelled by similar goals and assumptions, and at times hampered
or shaped by differing ones. In addition, cognition appears to have played a role
when the embedded humanist participated in a change in the direction of the labo-
ratory research.
210 Erik Fisher and Roop L. Mahajan
Part I: Humanistic Engineering in an Engineering College
An Engineering Art Gallery

From 1998 to 2003, as the humanities advisor of the College of Engineering and
Applied Science, Fisher ran an array of interdisciplinary programs, both curricular and
extracurricular, including curating a small art gallery. The 500-square-foot Connec-
tions Gallery was situated just inside the main entrance of the college, whose imposing
brutalist architecture loomed on the east side of Boulder’s otherwise Spanish-tile-and-
flagstone-bedecked campus. With its mission to “connect art, engineering, and
society,” the gallery symbolized the notable interest of a handful of engineering
administrators to open up the college and its curriculum to an infusion of humanistic
ideas. Artistic installations, in the words of one associate dean, were intended to
introduce engineering students to “a different way of thinking.”
In 2002, Fisher and Mahajan, the newly appointed dean of the College of Engineer-
ing and Applied Science, met for the first time when they got together to discuss the
gallery. Mahajan, then a professor of mechanical engineering and a former AT&T Bell
Labs Fellow, had just accepted the appointment as interim dean. He was pleased to
learn not only that the college was showcasing its newly constructed art gallery, but
that someone without an engineering background was behind the endeavor. His
assumption had been that the college’s commitment to liberal and creative arts had
been minimal and that the humanities advisor had not made any meaningful efforts
to integrate the humanities and social sciences into the engineering curriculum. He
was determined to establish a clear line of communication and to ensure that the
image the gallery was projecting was in keeping with his intentions for the college as
a whole.
After an initial and somewhat tense exchange, the two of us quickly overcame our
various apprehensions. Our conversation turned more generally to the gallery’s
purpose and the role of the humanities in engineering education. We agreed that the
humanities could serve as a means for self-development, for posing critical questions,
and for broadening technical education in worthwhile ways. After our initial encoun-
ter, we began to explore these ideas in relation to the visual arts. The gallery hosted
exhibitions in a variety of media that explored social, cultural, and conceptual dimen-
sions of engineering.
One installation, sponsored by Ball Aerospace and entitled “Imaging the Invisible,”
began with the technological capability to render visual images of what is otherwise
undetected by the human eye—for example, landscapes covered with dense jungle
canopies, subterranean geological formations, stunning interstellar phenomena, and
Embedding the Humanities in Engineering 211
the activity of criminal underworlds and private life. The process of representing these
previously imperceptible worlds to the probing eye was not, according to the exhibit,
solely the product of satellite and aerospace engineering know-how. Rather, this
process entailed other forms of invisibility: the role that visual artists, graphic design-
ers, technical writers, and other actors play not only in supporting engineering work
but in animating its outputs with focus, shape, and vigor. Nor were decisions to gather
certain data or issues of its availability entirely separate from those who launched the
mechanical birds. Exhibits such as this one were meant to reveal to the inhabitants
of the college ways in which the aesthetic, societal, and ethical realms subtly and yet
powerfully permeate and are permeated by the technical.
Our interactions around this and other gallery activities gave way in due time to
new forums of exchange. We were soon engaged in conversations—with deans from
other colleges, department chairs, and federal agency funding officers—about giving
the arts, humanities, and social sciences broader roles within the college. These discus-
sions were shaped in large part by Mahajan’s vision for the college institution as a
whole, in which he sought to establish a group of synergistic research and education
initiatives—including microtechnology, nanotechnology, biotechnology, and earth
systems engineering—and aimed to distinguish the college intellectually and cultur-
ally in order to continue to attract both talent and resources to the Rocky Mountain
region.
Congruent Mental Models

Upon reflection, we had each separately hit upon the notion that interdisciplinary
activities, both cognitive and social, could enhance practical scientific undertakings—
whether by grasping and relating concepts or by creating new research pathways. This
common perspective can be characterized as a shared mental model (Gorman 2005).
In Mahajan’s case, his vision for the college was informed by values rooted in his
professional and personal experiences. At Bell Labs, he had developed the conviction
that interdisciplinary interaction facilitated a “creative reprieve” in which the inter-
mingling of ideas and perspectives provided relief from the tedium of technical details,
a relief that in turn allowed insights for new endeavors to emerge. There, he had
employed the phrase, “the buds bloom at the borders,” to suggest that different out-
looks, experiences, and even disciplinary training can give rise to worthwhile
exchanges. Meanwhile, Fisher, whose experience at St. John’s College had steeped him
in original scientific and philosophical Western texts, had been experimenting with
courses that conjoined such subjects as philosophy and mathematics, science and
literature, and engineering design and writing composition. In the process he had
discovered that open-ended and conceptual inquiry in technical settings, in addition

to being a pleasurable end in itself, could lend value to engineering education in
unanticipated ways. Classroom discussions took on new relevance for his students,
some of whom regularly reported going about their engineering coursework more
successfully as a result of grasping and comparing fundamental mathematical and
philosophical concepts.
We also shared concerns about the unintended consequences, emergent effects,
and increasing roles of technology in society. Mahajan had a keen sense that recent
advances in technology had been buoyed and advanced by expectations and promises
that were not readily forthcoming. Increased leisure time, for instance, had been one
expectation behind the ever faster and cheaper computing machines made possible
by the invention of the transistor; in reality, the trend toward higher productivity
makes it hard to see how our technological advances liberate rather than enslave users.
This observation had long ago sparked Mahajan’s sense that engineers—like a collec-
tive magician’s apprentice—had played a role in giving rise to “unfettered technol-
ogy.” Fisher’s background in classical political philosophy in turn led him to question
the prudent governance of modern knowledge production. He found discomfiting
correlations between the lack of distinctions his engineering students made among
the concepts of information, knowledge, and practical wisdom and the institutional
arrangements for allocating knowledge resources and social responsibility.
In short, we each enjoyed cultivating the practical utility of higher pursuits such
as learning, and yet cast a critical eye upon the unreflective employment of
knowledge.
Humanistic Engineering
Over time, we envisioned a vibrant and successful engineering education program that
combined solid scientific foundations, hands-on design and research experience, and
lively humanistic inquiry—that was, at the same time, sensitive to the practical, unin-
tended consequences of its exuberance. Our efforts were focused on the undergraduate
curriculum, and we aimed to build upon existing programs that Fisher had created
which emphasized the development of a self-critical capacity in relation to technical
modes of thought. To fulfill our goal of “techno-humanistic” integration in the engi-
neering curriculum, we sought to include liberal arts components in the technical
curriculum and to make a more compelling case to students for their inclusion (Fisher
and Mahajan 2003).
We used the term “humanistic engineering” to describe our new program and the
skills and ideals that characterized it. Besides suggesting a critical and reflective quality
that we believed engineers had to possess in order to better grasp the ethical, human,
and social dimensions of their work, the phrase also implied a creativity and sharpness
of mind that we felt would increase the practical value of their work in the process.
We imagined future engineers who could draw from and supplement existing para-
digms in engineering and the humanities to develop more context sensitive solutions;
as we wrote: “Humanistic Engineers … are able to initiate and engage in effective
dialogue with non-technical audiences regarding socio-humanistic critiques of engi-
neering processes and products and … perform their own socio-humanistic critiques
in the absence of such dialogue” (Fisher and Mahajan 2003). This statement tasks
engineers and technical practitioners—and not only regulators, members of the public,
and experts in fields other than engineering—with the job of initiating integrative
efforts and of generating critical insights into their work. From our perspective, engi-
neers would need to be able to think through and address the social and ethical
dimensions of their work, regardless of whether others helped them in this task, so as
to frame problems and generate solutions in new ways. It is not that we failed to see
a role for civil society, professional groups, and other disciplinary practitioners; rather,
we held that it often would not be feasible to bring together such actors with engi-
neering researchers, designers, and decision makers at key times in order to produce
more informed and efficacious critiques. For, while there are pressing reasons to focus
on the broader social worlds, groups, and discourses that surround technology in
society, our experience and professional commitments as engineering educators had
led us to take a different tack: to address the intellectual activities that already ani-
mated the earliest institutionalized processes of technology in the making.
The term “humanistic engineering” was an attempt to fuse, on the one hand, the
sense of critical insight and discovery that can proceed from the practice of open-
ended inquiry associated with the humanities; and on the other, the employment of
this in the service of engineering innovation that is optimally directed at socially
relevant needs, but in a context-sensitive manner. Despite our shared terminology,
our conceptions of the humanities and of engineering were not identical. We had
different understandings of the nature of both fields, and thus envisioned directing
their employment toward subtly different ends. Mahajan, who had studied philosophy
as an undergraduate, placed tremendous value on the activities of humanistic reflec-
tion, and saw the place of the humanities in engineering—at least initially—as provid-
ing a creative stimulus for new ideas to emerge. Fisher, on the other hand, was
committed to the view that humanistic inquiry, as an attempt to grasp wholes and
reveal foundations, could actually lead more directly to fundamental insight and
self-discovery.
Dialogues between Two Cultures

A central forum in which we and others explored the intellectual and cultural ground-
work of humanistic engineering was Fisher’s faculty seminar project, Dialogues
between Two Cultures, supported by the National Endowment for the Humanities.1
The yearlong seminar and lecture series brought together a core group of engineering,
natural sciences, humanities, and social science faculty. The group met twice monthly
to discuss core readings with invited interdisciplinary scholars. Monthly themes (and
visitors) included art and technology (Natalie Jeremijenko), science and religion
(Arthur Zajonc), and social networks and trading zones (Michael Gorman), among
others. The project allowed participants to engage in regular, sustained, and focused
conversations among people who held radically different disciplinary perspectives.
Several participants went on to design and teach collaborative interdisciplinary courses
or to pursue interdisciplinary research proposals.
One of the functions of the seminar series was to explore potential areas of con-
vergence between the humanities and social sciences with their disciplinary counter-
parts, as well as clear divergences. In a seminar on the so-called “science wars,”
Mahajan articulated a role for the cognitive dimensions of science and engineering:
“Among scientists and technologists or engineers, many of the discoveries and innova-
tions are intuitive, and then we present evidence later on. … That is a wonderful
process. Now that intuitive thing is not so much evidence-based, it is not so much
experiments-based.” Not every seminar participant shared this view, and at least one
felt it adhered to an outdated distinction in the philosophy of science between dis-
covery and justification.2 That said, Mahajan’s emphasis on inductive insight over
deductive ratiocination, on intuition as a creative principle, ostensibly opens the heart
of scientific discovery and engineering design (and hence education) to a particular
form of empirically based humanistic inquiry. His distinction—reminiscent of the
“reprieve” that gives rise to creative activity—between “intuitive” content and “ratio-
nal” structures was commensurate with Fisher’s interest in cognitive and social
processes—writing composition, product design, decision making—in which an inde-
terminate path is taken by actors through otherwise discrete steps, stages, or categories
(Fisher and Benassi 2003). In short, we were each sensitive to the role of ideation
in stimulating innovative pathways within formal structures, both conceptual and
institutional.
The prospect of facilitating fruitful exchanges between the critical cultures of the
social sciences and humanities and the productive cultures of science and engineering
ran throughout the Dialogues project. In the same seminar session, possible connec-
tions between modes of thought and the larger workings of science can be seen in the
following separate remarks, by Mark Winokur (Department of English):
At its best for me, interpretation creates—recreates—the student as a kind of paranoiac. Question-
ing everything about the world, not just literature, not just interpreting literature but, the hope
is, beyond literature, the student will be asking questions about the way the world works includ-
ing, hopefully, the way that science works, the way that every discipline works, the way in which
one perceives the universe to work, and certainly one’s own self.
And by Clayton Lewis (Department of Computer Engineering):
My sense is that the best scientists … are the ones that are in fact skeptical of themselves, of
what they do. Here again, I was most impressed by those articles [in Labinger and Collins 2001]
in which there’s a kind of assumption that really the greatest and most acute criticism comes
from within the community, not from without.
Despite the different conceptions of skepticism and of the self that these two
statements suggest, they both link the cognitive aspects of interpretation and criti-
cism to a more robust understanding of the “working” and “doing” of science.
Lewis’s emphasis on scientific self-criticism suggests a form of scientific autonomy
that would include an ability to take into account and productively benefit from
internally generated critiques. Winokur, on the other hand, emphasizes a form of
training in mental habits—in this case that of literary interpretation—that suggests
a skill consonant with Lewis’s picture, but nonetheless originating from outside of
scientific practice per se. These two conceptions, of critical capacity that rests within
versus outside of science and engineering, emerged in different guises throughout
the Dialogues series. We return to these ideas below, where we reflect on the notion
of humanistic engineering and explore different forms of interdisciplinary trading
zones.
Conversations such as the two-hour Dialogues seminars, partly nested within the
larger institutional context of an engineering college, foreshadowed our further col-
laborations. As a humanist formally embedded within an engineering college, Fisher
had created a space for sustained social and intellectual interactions among various
disciplinary cultures—a workshop of sorts, which constituted a type of laboratory of
its own. Soon, Fisher’s interest in interdisciplinary inquiry conducted in small groups
and Mahajan’s interest in institutional transformation came to provide additional
opportunities for our collaborative interactions that would eventually find their way
into an engineering research laboratory. For the next phase of our undertaking was
an attempt to bring humanistic engineering to the level of graduate engineering edu-
cation and research.
Part II: An Embedded Humanist in an Engineering Laboratory
An Embedded Humanist
In September 2003, Fisher accepted the invitation to formally join Mahajan’s Thermal
and Nanotechnology Laboratory in the Department of Mechanical Engineering.
Shortly thereafter, he began referring to himself as an “embedded humanist.” He
coined the term in a somewhat lighthearted attempt to explain to his colleagues what
a scholar like him was doing in a laboratory. Mahajan soon also began using the term
to explain his objectives in reconfiguring his lab to other laboratory directors, some
of whom resisted the term because they associated it with “embedded journalist.”
However, unlike an embedded journalist, who reports to an audience outside the
environment that he or she has joined, Fisher’s audience was inside the environment
he had entered. During the thirty-three-month experience, Fisher interacted with
numerous researchers and laboratory personnel, followed a variety of laboratory proj-
ects, and conducted several studies. As a member of the research group, he had desk
space; attended weekly meetings; participated in equipment training sessions, events,
and professional conferences; and interacted informally with the researchers through
his regular presence in the lab, its offices, and its connected Nanoscale Fabrication
and Characterization Laboratory. As a working member of the laboratory, he also made
regular presentations to the rest of the group on his research progress. Over time,
Fisher gained a wealth of interactional abilities, which formed some of the basis for
what later became tangible contributions to laboratory research practices.
Integration Policy
In December of 2003, just a few months after Fisher joined Mahajan’s laboratory, the
Twenty-first Century Nanotechnology Research and Development Act of 2003 (NRDA),
which required “societal concerns … [to be] considered during the development of
nanotechnology” by “insofar as possible, integrating research on societal, ethical, and
environmental concerns with nanotechnology research and development,” became
public law (U.S. Congress 2003). The legislation placed repeated emphasis on inter-
disciplinarity and administrative coordination, two aspects that can logically be
applied to the sociotechnical integration it called for (Fisher and Mahajan 2006a). The
act was also largely unprecedented and went beyond traditional policy models (Bennett
and Sarewitz 2006; Fisher and Mahajan 2006a; Macnaghten, Kearnes, and Wynne
2005). Policies that bring societal considerations to bear on technological trajectories
tend to occur on one side (upstream) of research and development activities, in the
form of research policies, or on the other (downstream), in the form of regulations.
One of the things that made the NRDA unique as a science policy prescription was
its inclusion of midstream research and development activities as a legitimate site for
interdisciplinary interaction aimed at the responsible development of science and
technology (Barben et al. 2008; Fisher, Mahajan, and Mitcham 2006).
Given the work on nanotechnology that had recently begun in Mahajan’s lab, and
our interest in ultimately developing humanistic engineering within a context of engi-
neering practice, we chose to frame our lab-based collaborations in terms of this larger
policy development. For Mahajan, the NRDA validated the need to lay the foundations
for a graduate engineering program in humanistic engineering; for Fisher, it was an
opportunity to conduct empirical studies into central intellectual and policy chal-
lenges facing modern industrial societies—such as the negotiation of human values
and material practices—in the unique setting of laboratory knowledge production.
Sociotechnical Integration
The terms “sociotechnical integration” and “integration of science and society” served
us sometimes as a means and sometimes as an end. While we both had been using the
term “integration” for several years (e.g., Neeley et al. 2002), although in differing con-
texts and meanings, the word gained prominence after it appeared in the NRDA. The
NRDA coincided with our shift from using the concept of “techno-humanistic integra-
tion,” which was closely linked to our interest in the cognitive aspects of interdiscipli-
narity discussed above, to using “sociotechnical integration” (Fisher and Mahajan
2006a), which was oriented toward broader social goals and institutional contexts and
thus took sociological findings (e.g., Bijker 1995) more explicitly into account.
Notably, the legislation’s language called for integrating research, not necessarily
researchers. Certainly, there is a place for researchers to perform this work and hence
for interdisciplinary interactions to embody much of the integration. Our interest, if
not the goal of the NRDA, in exploring the integration of disciplinary approaches in
the laboratory was less for the sake of finding new social arrangements in themselves,
however, than for understanding and pioneering new knowledge generation tech-
niques, safeguards, and outcomes. New forms of social epistemology, and not merely
enlightened engineers, may indeed be necessary to bring these about—but we were
interested in how much could be asked of engineers in the absence of such new social
arrangements.
Seamless Integration
In 1989, Mahajan had received a Bell Labs Fellow Award in recognition of his work
on the seamless integration of fundamental and applied research. Accordingly, we
attempted to build on our interdisciplinary perspectives as well as the confluence of

theory and practice in order to get as close as possible to new forms of knowledge
production. The term “seamless integration of science and society” came to symbolize
our attempts to build upon past research accomplishments and to probe the limits of
a new practice that might hold similar utility. “Seamless integration” implied that the
elements to be integrated were not treated separately but were coupled frequently and
at key junctures, so that the elements influenced one another and so that the coupling
occurred at every stage of an engineering project. Seamless integration also served as
a baseline posited to assess how far we could actually come and to assist similar
attempts in the future.
We eventually abandoned the idea of seamlessly integrating sociohumanistic and
technical considerations at every conceivable stage of an engineering research project.
The decision was due in part to time constraints and in part to a study by Fisher that
revealed differing stages of laboratory research projects—which implied differing pos-
sibilities for and constraints upon integration. The shift also occurred in conjunction
with a more explicit focus on processes. As we wrote,
Ideally, seamless integration would introduce broader considerations at every stage of a research
project and make more explicit the extent to which such considerations are already implicitly
present. It would require a tendency to remain open to questions and concerns and a willingness
to revisit them at later stages. Due to practical constraints of engineering research, however,
integration will usually be limited to opportunities for subtle modulation of ongoing processes.
(Fisher and Mahajan 2006b)
Seamless integration thus set the stage for the more modest but demonstrably potent
approach of midstream modulation.
Midstream Modulation
Mahajan desired a replicable methodology and encouraged Fisher to develop a “pro-
tocol.” Although sympathetic, Fisher had reservations about the limitations of an
overly reductive and deterministic approach. After initial attempts to formulate the
basis for a prescriptive process, his doubts were soon intensified by the resistance of
several laboratory colleagues. When he modified his initial efforts, it became clear that
we did not have the same understanding of the implementation tool for sociotechnical
integration. This can be indicated by our different understandings of the term “pro-
tocol”: while Mahajan’s reference point was from medical research, and involved a
normative flow chart to address ethical harms and risks, Fisher preferred to associate
the term with an interview schedule such as that used in sociological and ethnographic
research. The differences in our conceptions were not clear at first, which caused
confusion. When they did become clear, Mahajan’s desire for an instrument that
would facilitate normative prescription contrasted with Fisher’s desire for a richly
descriptive basis for integrative laboratory research practice that was in accordance
with his ethnomethodological observations. For instance, Fisher had come to appreci-
ate that the integration of societal and technical considerations was in fact already
occurring, and that the problem to be worked on was that researchers were largely
unaware of their own role in such de facto integration.
It is worth noting that our collaboration could have taken any of a number of dif-
ferent forms at this point. In short, we were encountering the problem of incommen-
surability. Despite a number of overlapping values and synergistic ideas, we had
different expectations and assumptions about the nature and role of ethical questions
in lab research and about what kind of social and cognitive actions would constitute
a productive response to these questions. The differences were no doubt due in some
part to the different research paradigms we employed and had observed others employ.
Fortunately, we were able to find a way to sidestep, if not surmount, this initial
incommensurability.
As a result of our diverging expectations, and still keeping in mind the NRDA
language, Fisher developed a framework for midstream modulation that would
satisfy our competing goals. It consisted of a three-stage dialectical process that
modeled sociotechnical integration and change. The process factored in both descrip-
tions of existing material practices (de facto modulation) and an awareness of them
(reflexive modulation) as a cognitive precondition for acting upon normative pre-
scriptions (deliberate modulation). Using this framework and Fisher’s observations
of laboratory decision making, the laboratory researchers held group and individual
meetings to further develop a descriptive model that could apply generically to
any laboratory research decision. The research group thus negotiated in various
iterations with Fisher to develop the instrument that was meant to capture oppor-
tunities for integration and modulation. The output of this process has elsewhere
been termed an “ethnographic invention” (Fisher 2007). Here, we note that the
negotiations that gave rise to it are reminiscent of those in trading zones designed
to either work around or overcome paradigmatic differences. The new protocol
thus converted what was a top-down directive into a bottom-up cultural artifact
that served as the basis for a new interview schedule. Moreover, the new interview
process itself could now take into account—if not facilitate—intentionality and
normative goals.
In the spring of 2006, toward the end of the thirty-three-month laboratory experi-
ence, Fisher conducted a twelve-week field study in which he interacted with graduate
researchers to assess the feasibility of sociotechnical integration in a laboratory context.

The results suggested that midstream modulation activities were able to enhance if
not stimulate graduate engineering researchers’ awareness of cognitive, social, and
material research-conditioning factors (Fisher and Mahajan 2006b). Moreover, in a
few cases, where there had been a high level of interdisciplinary engagement, this
awareness correlated to specific and concrete changes in research decisions, thus influ-
encing the direction of engineering research (ibid.). As stated elsewhere, “While the
protocol rendered elements of [one engineering researcher’s] decision making more
visible to [the embedded humanist], it also made them more visible to [the engineer,
who demonstrated a] reflexive capacity to identify and align ‘social’ considerations
with ‘technical’ alternatives in order to solve a complex problem” (Fisher 2007). More-
over, the research practices themselves underwent several changes that were correlated
to the engagement, and specifically to social and environmental concerns that were
articulated during the interactions. These changes in research practice—which included
altering the disposal method and chemical synthesis techniques for carbon
nanotubes—were considerably more benign in terms of their environmental and
human health dimensions.
Meanwhile, the midstream modulation experiment did not disrupt laboratory
conduct or compromise its outputs. If anything, it aided the research—as illustrated,
for example, by one researcher who noted on several occasions that discussing his
research with Fisher helped to clarify his own thinking about his work. As a proof of
concept, then, the collaborative engagement indicated that the goals of humanistic
engineering are reasonable and worthwhile. Not only are such interactive methods
consonant with academic engineering research practices, it appears that they could
be applied to other engineering activities, such as design, and to other forms of lab-
based science. The techniques and concepts used in the experiment, which were
primarily aimed at asking questions, thus had dual value both for enhancing the
creativity and productivity of research and in integrating social and ethical consider-
ations into the research.3
Part III: Reflections on Trading Zones and Interactional Expertise
Collins, Evans, and Gorman (2007) produced a framework to relate different types of
trading zones (Galison 1997) and to chart their evolution, and which incorporates
Collins and Evans’s (2002) notion of interactional expertise. The framework’s calculus
of characterizing a trading zone by means of its end state can be instructively applied
to the above two-part account in several respects.
Degree of Integration in Trading Zones

The fourfold characterization of trading zones derives from a two-by-two matrix that
identifies zones in terms of their placement along two axes: one axis represents the
degree of coercion or collaboration involved in the interaction; the other, the degree
of heterogeneity or homogeneity of the end state of collaboration (see figure 2.2). If
we consider this latter axis, we find there are indeed differing degrees of homogeneity
and heterogeneity in our conceptions of humanistic engineering and in our attempts
to enact it—particularly in the case of embedding a humanist in an engineering
laboratory.
Applied to our ideas about and work on humanistic engineering, this model sug-
gests that we actually envisioned a pair of sequential trading zones: while the first one
comprises the educational environment and training ground for humanistic engineer-
ing, the second one denotes its practice. Humanistic engineering education is made
up of an interdisciplinary combination of pedagogical goals and approaches, and
consists of a relatively heterogeneous state in which the disciplinary cultures involved
have a high degree of collaboration and yet remain distinct and distinguishable. This
stage of interdisciplinary collaboration, which Collins, Evans, and Gorman term a
“fractionated trading zone,” can be likened to both the provocative situating of an art
gallery within an engineering college and the educational goal of producing human-
istic engineers.
In Mark Winokur’s example during the Dialogues project of a highly skeptical and
inquisitive literary critic who is able to turn her attention to the workings of science
qua science (as opposed to science qua literature), the requisite skill set and training
in literary criticism would most likely occur in a classroom environment in which the
disciplinary profile of English, comparative literature, classics, etc., is highly visible—
even if it has integrated elements of, say, civil or mechanical engineering. The same
would be true for most analytical mathematic and scientific skills—even if they
included elements of philosophy, sociology, psychology, and political science. One
imagines that linguistic innovations would indeed occur as the pedagogy of this envi-
sioned program developed and matured, but the traditional quantitative and qualita-
tive reasoning skills associated with engineering and the humanities, respectively,
would likely remain as distinctly characteristic as the separation between the qua-
drivium and the trivium in long-standing conceptions of the liberal arts—as would
the identity of their primary classroom instructors and other facilitators, as a designa-
tion like “humanities advisor” suggests.
Humanistic engineering practice, on the other hand, consists of a synthesis of
culturally segregated skills, competencies, and habits of thought that, while they can
be decomposed into humanistic and engineering approaches, have become largely

indistinguishable from one another. This “interlanguage trading zone” represents a
homogeneous end state that is in keeping with Clayton Lewis’s notion of criticism
emerging from “within the [scientific] community.” It is also clear in the description
of humanistic engineers who are able to “perform their own socio-humanistic cri-
tiques” (Fisher and Mahajan 2003) in the absence of dialogical relations with practi-
tioners or decision makers in fields other than engineering or with lay citizens. While
an interdisciplinary basis would in theory allow the humanistic engineer to tap dif-
ferent disciplinary techniques and modes of thought, the modes of thought them-
selves are sewn up in the identity and mental habits of the practitioner.
We also note that during our first year of collaboration our own interactions
revolved largely around shared terms and ideas, several of which we had generated
relatively early on. We became conversant in each other’s languages and cultural
contexts and began to develop our own shared terms or “jargon.” We also tried to
downplay and, in some cases, to transcend disciplinary and cultural distinctions,
which nevertheless are evident in hindsight.
The phrase “humanistic engineering,” for instance, symbolized both the creativity
that Mahajan was interested in and the self-critical inquiry Fisher was focused on.
Because our interests overlapped in this way, we were friendly to reflection and to the
time it takes to engage in it—a crucial factor that helped protect our unorthodox col-
laborations from mainstream skeptical reactions. In particular, the phrase served to
establish an interactive niche between Mahajan, who tended to emphasize the human
and cultural development of the engineer, and Fisher, who was more animated by
philosophical questions and the political implications of sociomaterial productions.
While the word “engineering” in this phrase served to identify the shared topical site
of action and observation, the word “humanistic” afforded a fruitful ambiguity, in
that it at once signified two anticipated outcomes of reflective inquiry by technological
actors: the informed development of the self, and the construction of the sociotechni-
cal world tempered by humility (cf. Jasanoff 2007). At any rate, the phrase exemplifies
the notion of a fractionated trading zone in which language allows productive inter-
change even when the meaning of terms can be multiple.
Evolution of Trading Zones

The heterogeneity of a trading zone, which can in our case be thought of in terms of
external versus internal critique, not only represents two different forms of interdis-
ciplinarity but describes a general process of evolution—for instance, from education
and training to practice. In our case, however, as our collaboration coalesced around
a specific challenge and our attempts to implement humanistic engineering moved

from undergraduate to graduate engineering education, our conceptions of what we
were doing appear to have taken heterogeneity more into account as time went
on—even as we sought to blur disciplinary distinctions.
While the humanistic engineer embodies a state of disciplinary homogeneity, the
embedded humanist calls attention to heterogeneity. Arguably, the notion of embed-
dedness contains the notion of an “other,” regardless of how successfully the other
has been incorporated within a new culture; for it remains nonnative, a transfer, a
foreigner. Thus, as we sought to implement our program of humanistic engineering
within the context of laboratory practice, we moved from the guiding motif of a highly
integrated form of interdisciplinarity to the more heterogeneous state of an outsider
within the highly bounded space of an academic laboratory. The shift to a more pro-
nounced fractionated trading zone was perhaps in part due to the different contextual
demands and the shift from program development to research formulation and
conduct.
We have noted the association of our concept to the embedded journalists deployed
by the U.S. armed forces during the 2003 invasion of Iraq. As we suggested earlier,
there are key differences: embedded journalists are intended to relay information to
those outside of a unit, without affecting its internal operations; an embedded human-
ist, on the other hand, seeks to learn through intervening to what extent the internal
routines of the laboratory may undergo and tolerate change as a result of self-critical
activities. The term “embedded humanist” was also meant to stress the novelty of the
enterprise in light of more traditional participant observation: ours was an attempt to
interact and collaborate in order to issue policy recommendations on engineering
education that were inspired from outside of engineering education and yet that
would be grounded in firsthand experience with it. Unlike the self-sufficient and
enlightened engineer that we had formerly conjured up, the work of the embedded
humanist requires a synergistic dependence of dissimilar practitioners on one another.
In that this work involves sustained engagement of laboratory practitioners, the
embedded humanist is in a position to become part of the convergence of goals,
strategies, and sociomaterial configurations. Yet, despite the fusing of practice and the
intermingling of knowledge, the collaboration retained its interdisciplinary heteroge-
neity and the volitional autonomy of its members. We find little evidence of co-
optation of either the embedded humanist’s research or that of the laboratory
practitioners, and we note that some of the key decisions made by both sides of the
various collaborative partnerships were independent of the understanding and expec-
tation of the other.
The shift from disciplinary integration to differentiation can also be seen in our
own roles. As the embedded humanist forged greater relationships with those who
occupied the laboratory on a daily and hourly basis, he identified more closely than
before with their bench science and came to better understand their perspective of
working under the institutional figure of the laboratory director.
From Coercion to Collaboration

The evolution of our collaborations in this first year went from a somewhat enforced
meeting in which we embodied distinct formal and institutional roles, to a collab-
orative set of ideas and activities in which we generated, responded to, and attempted
to align broader visions and strategies, existing opportunities and resources, and
local means and venues—a highly collaborative trading zone in which we merged
language, practice and to some extent identity. While the institutional structures
that defined our roles initially exerted a coercive influence, our interest in redefining
these structures and our ability to do so helped us open up additional possibilities
for others to then engage in unconventional interactions that were nested within
the broader and changing institutional structure of the College of Engineering and
Applied Science.
The development of a protocol for midstream modulation was also born of coer-
cion, albeit of a more gentle kind. Mahajan’s interest in a protocol that would extend
the notion of seamless integration from that of basic-applied to that of sociotechnical
integration was modeled on medical research protocols in which primarily expert-
determined and quantifiable risks were at play. But in the case of emerging technolo-
gies such as nanotechnology—and the early stages of innovation in general—where
risks are more obviously characterized by high degrees of uncertainty, ambivalence,
and potential contestation, Fisher resisted bounding research practices by clear markers
and prescriptive norms. And yet we were both normatively inspired to introduce
broader reflection into these practices. The close and ongoing observation of labora-
tory decisions that the midstream modulation decision protocol allowed, coupled with
the more modest interventionist notion that comes with the notion of modulation
(Rip 1998), enabled us to bring about not only a process but productive outcomes that
satisfied both Fisher’s critical and Mahajan’s pragmatic objectives. Without Mahajan’s
insistence on the normative applicability of the protocol, and Fisher’s insistence on
its empirical validity and sensitivity to context—including the laboratory researchers’
acceptance of our ideas—we might not have come up with an instrument with poten-
tial value from multiple perspectives.
Cultural Subversion?
It is valid to ask whether cultural subversion, in which “one party’s language over-
whelms that of the other” (Collins, Evans, and Gorman 2007), occurred in the experi-
ence of the embedded humanist. Although we find limited evidence of an enforced
trading zone, it might appear that the embedded humanist lost his identity in the
technical language, institutional arrangements, or material practices of the laboratory
(not to mention the steady uptake of the social sciences into the mix). Two observa-
tions caution against overly hasty conclusions that such forms of cultural subversion
did occur here, however. First, the negotiated decision protocol served as a template
that framed many of the collaborative interactions, especially when they became so
intensive as to require constant study and attention to scientific and technical details.
This conversational and analytical frame thus provided the embedded humanist with
a method for maintaining if not retrieving an observational distance and analytical
standpoint. Second, the role of the humanist was essentially to practice humanistic
inquiry—something no one in the laboratory did in a structured or ongoing way—
which resembles the kind of attitude and attention one might adopt in the case of
textual analysis. In other words, the laboratory became a topical site for humanistic
inquiry.
Ironically, Fisher had to repeatedly resist various efforts to cast him as the human-
ist spokesperson for “ethics” or to be the moral “conscience” of the group. Despite
encouragements for him to identify ethical concerns and to advocate ways to
address them, Fisher sought to tactically avoid taking outright prescriptive stances.
Instead, although he did not in most cases promote specific ethical or societal
concerns—or specific technical or scientific responses—he sought to enhance if not
stimulate reflexive awareness. Also ironically, Fisher made no conscious attempt to
alter research practices—which in itself may have been what allowed for practices
to in fact change.
The Role of the Embedded Humanist

In the final phase of the embedded humanist’s experience in the laboratory, what
began as observation of engineering research and researchers in the making shifted to
participation, and finally to collaboration. Fisher had to acquire what Collins and
Evans (2002) call interactional expertise if he was to understand and critically inquire
into the work that went on in the laboratory. He had to learn the language of
Mahajan’s research and the specific projects of a number of researchers he interviewed
and interacted with intensively. In fact, he managed to develop sufficient interactional
expertise to be mistaken for an engineer in a number of venues—including by a project

manager at a prominent national laboratory. Moreover, as time progressed, the engi-
neering researchers learned and adapted some of the language he employed. And yet
his interactions with the researchers were not limited to discursive ones: as a result of
these interactions, engineering research material practices underwent significant
changes, suggesting that Fisher had achieved a measure of contributory expertise.
Although the embedded humanist had a role in stimulating these changes in
research practice—changes that seem consonant with both Mahajan’s and Fisher’s
shared perspectives on the practical value of cognitive activity—Fisher did not dictate
the goals of the deliberate modulations, nor did he lobby for the alternatives that
would lead to their pursuit. By asking questions and “holding out” in them (Heidegger
1977), his role was to maintain cognitive tension and social space for negotiations to
emerge that otherwise might never have been opened up and amplified. Fisher con-
tributed to the project, then, in an indirect way, rather than through suggesting new
ideas or mediating the suggestions of others. Rather, he contributed through attempt-
ing to understand and articulate key social, material, and cognitive elements at work
in the system within which he was embedded. This is perhaps similar to the interpre-
tive work that Winokur described during the Dialogues project, in which questions
are posed and the tensions they evoke are kept alive in an attempt to conceptually
grasp a broader whole (in his provocative term, through paranoia). This activity thus
added value to the engineering research both by promoting the advance of the
research as such and by modestly controlling the undesirable effects of the research.
As an interactional expert, the embedded humanist had the role of constructing frac-
tionated trading zones.
A Suspended Trading Zone

One of the four types of trading zone that Collins, Evans, and Gorman identify is
characterized by high degrees of both collaboration and heterogeneity. They call this
a “fractionated trading zone” and identify two forms: “boundary object trading zones,
which are mediated by material culture largely in the absence of linguistic interchange,
and interactional expertise trading zones, which are mediated by language largely in
the absence of the material.”
An implication of their framework is that, if the trading zone is characterized by
both linguistic interchange and material practices, it is likely that the zone has devel-
oped into an interlanguage trading zone, one with high degrees of collaboration and
homogeneity. Yet there is also another possibility: a trading zone with high degrees
of collaboration and heterogeneity might continue to exist in that state without pro-
ducing a new culture, language, or science. We suggest that the lab-based interactions
of the embedded humanist unified an otherwise fractionated trading zone through
the shared jargon of the modulation protocol, the altered research decisions, and the
subsequent material configurations that developed. The embedded humanist thus
became part of the convergence of rituals, values, strategies, and material practices,
but retained important elements of heterogeneity. In occupying both sides of the
fractionated trading zone, then, which did not collapse into a homogeneous state, a
new type of collaboration can be identified. We describe this as a “suspended trading
zone.”
Cognitive Dimensions of Interdisciplinarity

The cognitive dimensions of interdisciplinary collaborations may be a worthwhile
element for understanding and characterizing suspended trading zones. In our case,
discovering shared mental models that linked reflection, innovation, and social
responsibility early in our collaboration and continuing to develop them over time
provided a strong basis for our interactions. This connection made it possible for
us to envision an intimate role for the humanities in both engineering education
and in engineering practice. Moreover, it likely played a role in the embedded
humanist’s involvement in laboratory research, which resulted in equipment rede-
sign, a new disposal method, and a new experimental technique. We attribute such
altered research decisions to the enhanced reflexive awareness of the engineering
researcher, which we see as a result of the midstream modulation activities. In
short, the embedded humanist was able to facilitate responsible development through
simultaneous reflection on ethical values and on alternative material pathways,
precisely because he did not advocate for either one. The change in engineering
research, he explains,
was instrumentally triggered not by the interjection of mandates or prescriptions, but by [the
engineer’s] own cognitive work of reflection, association, and invention. Rather than introduce
social or ethical considerations, the protocol instead allowed [the engineer’s] latent concerns to
surface. As an intervention, this engagement of research capacity was productive because of the
work of the subject—the engagement may have influenced practice, but to do so it required the
practitioner’s desire to remedy a perceived deficiency. [The engineer’s] recognition was, in turn,
enabled by [the humanist’s] ongoing attentiveness to his unfolding account of social processes
and material properties. (Fisher 2007)
The embedded humanist’s work thus consisted of observation, reflecting observations

back into the research context, and sparking inquiry about potential value-material
relations that emerged as a result. The central role of cognition in this interdisciplinary
collaboration may help explain how both sides of the suspended trading zone were
able to function without producing a new homogeneous state.
Arguably, there was some form of a shared mental model between the humanist
and the engineers in this case, in that they both employed the same analytical deci-
sion categories during their conversations and, it appears, in their cogitations. There
was also something of a shared goal of research, at least at certain moments, when
both humanist and engineer seemed focused on the question of to what extent socio-
technical integration was possible. Beyond this, it is difficult to specify whether there
was much, if any, shared cognitive content. The surfactant that maintained the col-
laborative enterprise was, in our view, rooted in a learning process—that of semistruc-
tured, ongoing, shared reflection and inquiry into evolving human practices (Rabinow
and Bennett 2009).
Conclusion
This chapter describes our attempts to embed the humanities into engineering in two
main venues: the physical, institutional, and intellectual spaces of an engineering
college and of an engineering research laboratory. We used the combined metaphors
of trading zones and interactional expertise to analyze these attempts. Our analysis
highlights a tension between two types of trading zone implicit in our conception of
humanistic engineering: a relatively homogeneous state of interdisciplinarity that we
initially desired and that Collins, Evans, and Gorman term an “interlanguage trading
zone,” and a relatively heterogeneous state of interdisciplinarity that we came to rely
on and that they call a “fractionated trading zone.” This tension shows up both in
the case of humanistic engineering and in that of the embedded humanist. The analy-
sis also suggests that creative forms of collaboration can emerge out of initially coercive
conditions. Finally, we argue that there are other, nonlinguistic aspects to the inter-
disciplinary collaborations we were involved in that are also not fully attributable to
material boundary objects. We suggest that cognitive dimensions may be an under-
utilized factor in theorizing interdisciplinary collaborations, and that they may be
useful in accounting for what we term the “suspended” trading zone.
Notes
This material is based upon work supported by the National Science Foundation under Grants
No. 0531194 and 0849101 and the National Endowment for the Humanities under Award No.
ED-22408-02. The views, findings, conclusions, or recommendations expressed in this chapter
do not necessarily reflect those of the NSF or the NEH.
1. The project is documented at http://www.colorado.edu/engineering/dialogues/.
2. Alan Franklin (Department of Physics), for instance, responded: “I think this is a very old-
fashioned view in philosophy of science, namely the context of discovery and justification. In
the old days the context of discovery was regarded as the subject for psychology. … For example,
take the old myth—it doesn’t really matter where Kekule got his idea for the benzene ring because
he dreamed about snakes swallowing their tails. God knows what a Freudian would do with that,
but what matters is what evidence was then provided to argue that benzene had a ring
structure.”
3. The experiment came to serve as pilot study for the Socio-Technical Integration Research
(STIR) project, which embeds social and human scientists in twenty laboratories of numerous
kinds in ten countries on three continents (http://cns.asu.edu/stir/).
References
Barben, Daniel, Erik Fisher, Cynthia Selin, and David H. Guston. 2008. Anticipatory Governance
of Nanotechnology: Foresight, Engagement, and Integration. In The Handbook of Science and
Technology Studies, 3rd ed., ed. Edward J. Hackett, Olga Amsterdamska, Michael Lynch, and Judy
Wajcman, 979–1000. Cambridge, MA: MIT Press.
Bennett, I., and D. Sarewitz. 2006. Too Little, Too Late? Research Policies on the Societal Implica-
tions of Nanotechnology in the United States. Science as Culture 15 (4):309–325.
Bijker, W. E. 1995. Of Bicycles, Bakelites, and Bulbs: Toward a Theory of Sociotechnical Change.
Collins, H., and R. Evans. 2002. The Third Wave of Science Studies: Studies of Expertise and
Experience. Social Studies of Science 32 (2):235–296.
in History and Philosophy of Science 39 (1):657–666.
Fisher, E. 2007. Ethnographic Invention: Probing the Capacity of Laboratory Decisions.

NanoEthics 1 (2):155–165.
Fisher, E., and M. Benassi. 2003. Writing as Design: Integrating Composition and Product Design
in the Undergraduate Engineering Curriculum. American Society of Engineering Education
Annual Conference Proceedings, Nashville, Tennessee.
Fisher, E., and R. L. Mahajan. 2003. Humanistic Enhancement of Engineering: Liberalizing

the Technical Curriculum. International Conference on Engineering Education Proceedings,
Valencia, Spain.
Fisher, E., and R. L. Mahajan. 2006a. Contradictory Intent? U.S. Federal Legislation on Integrating
Societal Concerns into Nanotechnology Research and Development. Science and Public Policy 33
(1):5–16.
Fisher, E., and R. L. Mahajan. 2006b. Midstream Modulation of Nanotechnology Research in an

Academic Laboratory. Proceedings of International Mechanical Engineering Congress and Exposi-
tion 2006, Chicago, Paper no. IMECE2006-14790, 189–195.
Fisher, E., R. L. Mahajan, and C. Mitcham. 2006. Midstream Modulation of Technology: Gover-
nance from Within. Bulletin of Science, Technology and Society 26 (6):485–496.
Chicago Press.
Gorman, M. E. 2005. Levels of Expertise and Trading Zones: Combining Cognitive and Social
Approaches to Technology Studies. In Scientific and Technological Thinking, ed. M. E. Gorman
et al., 287–302. Mahwah, NJ: Lawrence Erlbaum Associates.
Heidegger, M. 1977. Modern Science, Metaphysics, and Mathematics. In Heidegger, Basic

Writings, 267–305. New York: Harper and Row.
Jasanoff, S. 2007. Technologies of Humility. Nature 450:33.
Labinger, Jay A., and Harry Collins, eds. 2001. The One Culture? A Conversation about Science.
Chicago: University of Chicago Press.
Macnaghten, P., M. Kearnes, and B. Wynne. 2005. Nanotechnology, Governance, and Public
Deliberation: What Role for the Social Sciences? Science Communication 27 (2):1–24.
Neeley, K., E. Fisher, E. Maslen, M. Piket-May, J. Prados, M. Cherno, J. K. Brown, M. Schoultz,

and R. Spizer. 2002. Integration as a Means to Excellence in Engineering Education and Practice.
Conference Reports, Liberal Studies and the Integrated Engineering Education of ABET 2000,
Charlottesville, Virginia.
Rabinow, P., and G. Bennett. 2009. Human Practices: Interfacing Three Modes of Collaboration.
In The Ethics of Protocells: Moral and Social Implications of Creating Life in the Laboratory, ed. M. A.
Bedau and E. C. Parke, 263–290. Cambridge, MA: MIT Press.
Rip, A. 1998. The Dancer and the Dance: Steering in/of Science and Technology. In Steering and
Effectiveness in a Developing Knowledge Society, ed. A. Rip. Utrecht: Uitgeverij Lemma BV, 27–50.
U.S. Congress. 2003. Twenty-first Century Nanotechnology Research and Development Act of
2003, Pub. L. 108-153, § 2, Dec. 3, 2003, 117 Stat. 1923.
11 Can Trading Zones and Interactional Expertise Benefit Business
Strategy?
Bolko von Oetinger
The Need for Change in Business
The constancy of change is a hallmark of business. Management spends most of its

time initiating numerous minor changes and a few dramatic ones, and hopefully these
changes are implemented. The emphasis here is on the word “hopefully,” since the
track record of success in regard to substantial changes is somewhat mixed.
The academic trading zone is a communication/collaboration framework intended
to overcome well-defined boundaries across different science disciplines. The approach
comprises processes (e.g., common language, collaborative activities, tacit knowledge,
mediating, values, and identities) and structures (various settings for interdisciplinary
organization). The notion of a trading zone is used throughout this chapter, but in
the narrow sense of a neutral field with a collaborative “interlanguage” approach, or
what Collins, Evans, and Gorman (2007) have specified as an “interlanguage trading
zone.” In this chapter, I will explore whether the concept and practice of academic
trading zones can be applied to business.
Is the Trading Zone an Effective Framework for Change in Business?
In the realm of business, customers and competitors, and technologies and markets,
are always on the move. Firms must quickly adapt, often in a reactive way but prefer-
ably proactively. Therefore, identifying weak signals early on, evaluating their impor-
tance, testing new ideas and reflecting on the situation in the light of perceived new
evidence should be part of an ongoing strategy process. Although the logic sounds
simple, in reality it is daunting.
Great organizations are great because of their strong, consistent performance.
Accordingly, strong performance leads to strong beliefs. Strong beliefs create strong
mental models, which can in turn end up as unwritten laws and taboos. This is where
232 Bolko von Oetinger
the problem begins. This is not to say that business success is a problem—managers
are paid to engineer economic success—but that the consequences of their success
could turn into a weakness that threatens the life of the company: the more successful
businesses are, the greater the risk of suppressing weak signals that seem to disturb
the (mental) business model being pursued. In other words, we have the belief that
what should not be true cannot be true.
Although firms are social entities, the traditional firm is profoundly asocial: it has
strong boundaries and is therefore something of a walled city. The notion of the “firm”
already indicates an entity that is solid and steadfast. Life within this corpus (the
corporation) is based upon an established culture of specific processes, beliefs, gover-
nance, laws, and taboos. The advantage is obvious: changes within the scope of the
guiding business model are somewhat routine and therefore done quickly. If it were
otherwise, continuous efficiency gains and continuous innovations could not be
achieved.
The more difficult type of change is that requiring massive adaptations of the gov-
erning business model. Indeed, the modern firm pays a high price for its closed culture:
dialog and innovations are hampered, there is steadfast resistance to fundamental
change, and there is more exploitation within the walls than there is exploration
outside them. Consequently, it should come as no surprise that large, successful
organizations, despite their increasing efficiency, have often faced the threat of
extinction.
Despite the natural resistance to substantial change, the changes that occur in busi-
ness can be brutal. They break organizations apart, result in the firing of leadership,
separate units from their homelands, lead to the sale of business divisions, and engen-
der mortal enemies. If one wishes to “depart,” one must “break away” first. We owe
to the great economist Joseph Schumpeter (1975, 82–85) the famous notion of “con-
structive destruction.” One must truly destroy a business in order to build a new one.
Occasionally such change is reminiscent of a butcher’s work.
Logically, given the difficulties of change, most great changes in business are taking
place not at the center but at its periphery (outside the walled city). These changes
move from the periphery ever closer to the old center, eating into its customer and
economic base as they go. Therefore, a high price is paid due to the inability to change;
or—on the positive side—the business’s inability to change pays off for the challenger
who ignores the guiding (mental) business model. He can prosper for a while in front
of the walled city, and no one behind the walls will really care.
The science of organizational behavior has failed to help business with the process
of change, because it narrowly defines the firm as a rigid contractual entity. Today,
Can Trading Zones and Interactional Expertise Benefit Business Strategy? 233
even in business, we observe movements into entities that are easier to access, such
as open-source movements (e.g., Linux), Web 2.0 organizations (e.g., Wikipedia), and
the deconstruction of the value chain leading to complex global supply chain net-
works. Therefore, the walls are coming down. Yet, although we enjoy a more open
world in business, the change processes fail to bring satisfaction.
Thus, the business world, with its well-separated organizations, bears a resemblance
to the scientific world with its great diversity of disciplines. However, can the frame-
work of academic trading zones be applied to business? Will we gain insights into the
processes of change in business by analyzing “interdisciplinary collaboration in science
and technology” (Gorman 2005)?
A trading zone is a specific framework developed in the social studies of science.
The relationship, which is found in a variety of academic disciplines, is certainly dif-
ferent from the environment of change in the individual firm. Scientific communities
enjoy a much greater degree of freedom (in time and content) than businesses do,
since the latter are more competitive but also more fragmented. Nevertheless, business
and the sciences share certain characteristics, for both are “cultures” (Gorman 2005).
The more one specializes for the sake of efficiency, the more likely it is that boundaries
will be created to hinder or impede communication.
Seemingly, trading zones have succeeded to an extent: there has been an increase
in cross-disciplines (such as nanotechnology, biochemistry, computer science and
service science), in the cross-disciplinary issues of industry (such as sustainability), in
the interdisciplinary arrangements of universities, and in cross-cultural solutions (such
as global health). All these ventures raise the question of how to manage the “cross-
ing” of boundaries.
Strategists who look deeper into the experience of “crossing boundaries” might
derive their most valuable finding from the trading zone concept. In order to perform
their demanding changes, businesses must continually cross well-established bound-
aries. How, then, do the sciences perform such a task?
Participants of the conference on Trading Zones, Interactional Expertise, and Inter-
disciplinary Collaboration, held in May 2005 in Phoenix, Arizona, referred to the
substantial boundary crossings not only in technologies (e.g., nanotechnology, nano-
ethics, and regulatory bodies) and sciences (e.g., history of sciences, computer science
and service science) but also in product development (e.g., the nuclear bomb,
radar, and fishing gear), distribution in underdeveloped countries (e.g., feminine
condoms), and services (e.g., climate forecasts, IBM, and systems of education). Thus,
there were many business issues on the table, representing complex organizational
structures, different cultures, and different mental models.
The interdisciplinarity of sciences in trading zones has certain characteristics that

merit closer examination from a business perspective:
1. No-man’s-land: a structured site for exchange,

2. Creating a common language of exchange,
3. Creating a culture of exchange beyond language,
4. Creating special structures for exchange,
5. Trading zones for business: open questions.
1 No-Man’s-Land: A Structured Site for Exchange
Anyone who attempts interdisciplinarity in science quickly arrives at a set of questions

such as the following: What are the boundaries of science? What is a science? What
is a new field? What happens at the boundary? Whereas what happens at the center
of science (in the lab) is well understood, what happens at the edge of the discipline
is not as easily recognized. I will argue that simply exchanging the word “science”
with the word “business” shows that such questions are equally relevant to the busi-
ness world. Perhaps it is because science and business share the same anthropologic
roots, since humans have always been on the move, crossing perceived borders.
The physical space of “trading” is most visible in the anthropological heritage of
the concept. Already by the late nineteenth and early twentieth centuries, the geog-
rapher Friedrich Ratzel and the sociologist Georg Simmel had made insightful observa-
tions regarding the spaces between tribes and/or countries. Ratzel introduced an
invaluable distinction between a “borderline” (which is exceedingly rare from a cul-
tural or anthropological standpoint) and a “border belt,” i.e., a “border region” (Ratzel
1899, 1:257–278). Most “borders” are not “solid lines” but elongated regions. These
“border regions” are culturally very rich, because populations from different areas
intermingle there. Typically, the great trading places are found in such regions. Georg
Simmel (1908, 703–704) used the notion of “empty space,” which he described as the
“deserted area” between two dominions; “I do not do something to you and you do
not do something to me.” He refers to “encounters … on neutral grounds” (ibid., 706).
You could move there and do things without being bound by your home rules. Ratzel’s
“border region” and Simmel’s “empty space” demonstrate the value of these neutral
grounds between two socially distinct units.
Collins, Evans, and Gorman (2007) have defined trading zones as situations in
which communities with deep communication problems manage to communicate. If
there is no problem in communication, there is simply trade but not a trading zone.
Such an assessment fits business well. Within an uncontested business model there is
always a lot of trade but no need for a trading zone. The situation becomes quite dif-
ferent if one is faced with substantial change.
Whereas many disruptive innovations in business end up in bitter fights between
the old and new business models (as seen in the airline industry), the trading zone
concept builds on deliberations (mediating and sharing). Whereas business is naturally
antagonistic or competitive (to put it diplomatically), the trading zone is by definition
a neutral area where antagonistic behavior ceases. Of course, there are times when
even the toughest competitors simultaneously collaborate and compete fiercely while
using neutral zones for their trade. One prominent example is the joint development
of the hybrid engine by Daimler, BMW, and GM. But such instances of cooperation
are not regarded as critical for future differentiation. In the example of the hybrid
engine, the car manufacturers keep their own development centers and will integrate
the new systems into their product range by keeping brand differentiation and techni-
cal product differentiation as competitive as possible. The less your competitor knows
about your future plans, the more you are advantaged.
In science the opposite is true. As a good scientist, you have to know your “com-
petitor’s” argument quite well and you argue with him in public. Perhaps the sciences
have found a common language to be the best approach in dealing with dramatic
differences. A notable exception in scientific competition are the laboratories that are
very careful about their equivalent of trade secrets—they do not share work on the
cutting edge until it is ready for publication, and they will have already secured any
patent rights they think they might be able to use to advantage.
2 Creating a Common Language
A key approach found in the trading zone is the development of a unique communi-
cation tool, creating a special (partial) language for the purpose of conducting trade.
In its simple form, such a “language” could consist of a few words, a “jargon,” while
higher forms would be a “pidgin” and finally a “creole,” or a full-fledged language
(Galison 1997). Although the concept of a trading zone language has been derived
from real trade experience and from real language, I use the concept of language and
trading zone in a purely metaphorical sense. The concept of having a specific language
implies that overcoming resistance (boundaries) is linked to cognition, and that cogni-
tion could be enhanced through better communication between the two areas.
The fact that one needs to communicate with distant (and therefore separate) busi-
ness areas exactly when one’s business is doing well, in order to identify potential
disruptions in the paradigm, seems to be a logical prerequisite for survival in business.

However, the siren songs of different business models suggest that businesspeople
should not confuse what are obviously apples with what are obviously oranges—which
is to say, one need not visit the other area because it seems inconsequential as an
addition to one’s business. Unfortunately, in business, the apples might metamor-
phose quickly into oranges. Therefore, one should constantly check for weak signals.
Successful businesses (with strong mental models) are usually also successful signal
suppressors. They refuse any communication with the “other,” which is unknown.
They confuse their mental model with reality. When somebody says to me, “The
reality of business is … ,” I know they are describing their own mental model.
A prominent example of this kind of communication between antagonistic busi-
ness models in the same industry is found in the airline business. A flag carrier (like
Delta or Lufthansa) is only able to understand the low-cost carrier (like Ryanair) if
the flag carrier accepts that its biggest asset (its highly protected hub) might also
become its biggest weakness. Flag carriers have grudgingly acknowledged (after con-
siderable price pressure from the low-cost carrier) that the alternate low-cost-carrier
business model (the point-to-point model instead of hub and spoke) has some eco-
nomic merits. Consequently, they have adopted that business model for some por-
tions of their portfolio. After having tried to fight the low-cost carrier for some time,
the flag carriers now try to absorb some of the low-cost carrier’s key success factors.
It is obvious that the flag carrier is not actively exchanging ideas (trading) with the
low-cost carrier, but is trying to imitate it. This is a type of exchange that does not
involve a trading zone, just observation, understanding the signals from customers,
and operational imitation. One has to study the details carefully to determine whether
there is a real trading zone and the gradual evolution of a new language in the indus-
try, or just a common language in the industry that already exists, making imitation
and adaptation easier.
A similar example comes from the attempt to develop food retailing as an e-busi-
ness. Here all business models failed, including that of Webvan, the most hailed
company of its time, with the notable exception of Tesco’s strategy to incorporate step
by step some of these ideas into their existing retailing structure. Again, there was no
trading zone, but Tesco used careful observation and transferred the successful factors
into its operations.
Often, businesspeople only vaguely realize that there is another specialized area
that could either be a threat or offer new opportunities. Because there is always some
noise around high-performing businesses, one can never be sure about what will
remain noise and what will become a strong signal. An example is the music industry’s
reaction to the emergence of the Internet. They were two separate, highly specialized
areas run by very different people: music moguls and Internet enthusiasts. The Internet
people, coming as challengers from the periphery, had not the slightest doubt that
music was another form of bits and bytes that could easily be distributed through
the Net. The producers at the music labels felt they had a legal barrier (copyright)
against any other distributor. Frankly, however, they were not engaged in selling
music, despite their claims; instead, their business model was the sale of CDs. They
had grown up in the “album business,” where one would hire an artist, pay her $20
million for the copyright, and sell her CDs, whereupon the investment could be
recaptured and the remainder kept as profit. The music moguls basically refused to
communicate with their enemies, the Internet enthusiasts, whom they regarded as
pirates. When the music industry realized the mounting threat, they called their
lawyers, choosing to fight the challengers rather than trying to absorb the essentials
of the Internet.
One interesting moment in the history of music on the Internet occurred when
one of the music giants (Bertelsmann BMG) bought an interest in Napster, the music
industry’s first Internet challenger. Was there a moment of collaboration in sight?
While Bertelsmann financed Napster, it also joined the legal actions of all other CD
producers against Napster. The so-called “collaborative” approach ultimately failed.
Apple Computer solved the problem for the consumer as well as for itself, but not
for the music industry. In developing the iPod, Apple understood that it had to deal
with the legal (payment) issues and with the Internet opportunities. It was reported
at the time that Apple’s CEO Steve Jobs had personally visited the major record label
companies in order to discuss the options, and that he was able to convey the neces-
sary trust in the deal. Did he find the “language” needed to talk to them in a way the
“computer geeks” could not?
The example of Apple’s iPod demonstrates that there was an agent of change who
was willing to actively reach out toward the alien territory, who was willing and able
to understand both areas, who developed a common proposal (a language) advancing
new opportunities derived from the Internet while protecting the copyrights of the
established system. This arrangement was as close as possible to a trading zone.
3 Creating a Culture of Exchange beyond Language
Is the creole—the specific language of the trading zone—more than a collection of

words? Can one have procedures without sharing the same values? Language is essen-
tially a cultural artifact, since it relates to history, implies ideas, presumes values, and
builds on tacit knowledge. As Lekelia Jenkins has demonstrated in the casework she
presented at the Phoenix conference (see chapter 8 in this volume), there are many
cultural factors that influence the language of trade, such as perceptions, communica-
tion habits, emotions, and institutions, that can turn the outcome in one direction
or another.
In Malinowski’s Argonauts of the Western Pacific (1922), the kula ring, the great gift
exchange of the Trobriand Islands, is described as a culturally shaped gift-trading zone
with its own great rituals, well-preserved ceremonies, strong beliefs, and expected
behaviors. Trading is a cultural act of people who interact socially. The kula ring offers
an array of trading conditions that are beyond economic utility: the islanders give and
take the gifts, independent of economic motives or assumed preferences. In fact, for
the Trobriand Islanders, the real trade and the gift exchange are two completely sepa-
rate activities, but they emphasize the importance of the noneconomic factors in good
economic relationships.
Given our interest in the creole as a language of exchange, it would be worthwhile
to know more about the culture of exchange that exists beyond language.
4 Creating Special Structures for Exchange
The complex communication process between different parties, well depicted by

Lekelia Jenkins and others in this volume—with its translational issues, hearing issues,
political climate, divergent personalities (ownership of ideas), participants relatively
more and less empowered, etc.—requires certain structures in order to give the lan-
guage an opportunity to emerge. These structures include, among other things, inter-
disciplinary meetings, technical and user committees, mediation groups, collaborative
work, interdisciplinary projects and teams, and interdisciplinary institutes. The overall
structure brings people closer together and helps align them with the “other” idea.
Consequently, the structure allows language to emerge.
This technique of using particular structures to absorb alien views is well known
in business. Such structures could include skunkworks, partnerships with other com-
panies, acquiring companies with different business models, job rotations, and changes
in the incentive systems or the exchange of personnel. It may sound trivial, but it is
still key: it is truly beneficial when people work together.
The transformation from first hearing an idea to instituting permanent change is
a complex pathway of self-discovery that needs further exploration. It brings to mind
a quotation attributed to Konrad Lorenz, although I was not able to verify its true
source. The English translation would be as follows:
Having said something does not mean it has been heard;

Having heard it does not mean having understood it;
Having understood it does not mean agreeing with it;
Agreeing with it does not mean holding on to it;
Holding on to it does not mean really applying it;
Applying it does not mean retaining it.
5 Trading Zones for Business: Open Questions
(a) In this paper I have distinguished between changes within an accepted business
model and changes that would modify the business model. The former are simple,
but the latter often constitute the real problem. It would be worth investigating
whether managing large changes might benefit from the concept of the trading
zone. The conference produced two arguments that could be of interest for further
research:
The content of change In some cases of disruptive innovation, as in the airline business
or in Internet-based food delivery, successful companies do not need a trading zone.
Instead, they can observe, imitate, and adapt if they are able to do so. In other cases
of fundamental change, like that in the music industry, it seems that the winners
(Apple, so far) get all the participating companies to agree on some economic terms
that would allow them to survive. Trading zones seem to be more likely to occur in
times of dissonance when everyone stands to lose unless they work together. Perhaps
the challenges facing the music industry at this point constituted a superordinate goal
(see Gorman and Werhane, chapter 12 in this volume), or perhaps the companies
were just responding to a prisoner’s dilemma situation—more research would have to
be done to be sure.
The financial crisis following the subprime mortgage collapse is another example.
Only if the financial institutions quickly develop new common standards will they be
able to master their destiny themselves; if they don’t, they will be regulated—that is,
they will be forced into a coercive trading zone.
In the future we might see more examples of these kinds of trading zones created
around environmental issues, which involve climate change, good governance,
sustainability, trust in business, etc.—all factors that cannot be solved by one single
player. Only if all players agree to certain standards can one do what is required of a
good citizen without giving one’s competitors a cost advantage.
However, if the dissonance favors (or is believed to favor) one party, then it may
suit that party to watch the others drown and to emerge as the leader afterward.
The process of change In business, big changes require strong leadership, because in
most cases one needs to “destroy” old beliefs. In the terms used by Collins, Evans,
and Gorman (2007) to express a more generalized model of trading zones, to make
big changes requires exercising sufficient “coercion” between seemingly “heteroge-
neous groups,” or a classical top-down “enforced” approach. By contrast, the creation
of an interlanguage is a “collaborative” procedure taking place at a lower hierarchical
level. The creole is a bottom-up approach that is used by people who are working
together. However, a certain homogeneity is engendered by investing in a joint lan-
guage. The process of learning parts of a foreign language and adapting them to one’s
own needs requires not only the acquisition of enough words and grammar but also
the acquisition of sufficient knowledge about the other culture. This is the impressive
part of the language: it contributes to some kind of homogeneity. Could this concept
of interlanguage be translated into business, or are we stuck with the model that is
enforced?
It is well known in business that the more and the earlier one engages people in
possible changes, the greater the potential for a self-discovered logic, which is in turn
the prerequisite of any robust change. That works well for the many minor, individual
changes that occur day to day, and it explains the enormous successes of continuous-
change programs like Sigma Six or employee suggestion systems, as long as people are
not disenfranchised in the process. The question remains open whether it is possible
to make the real, painful decisions through a collaborative, homogenizing approach.
On the one hand, the saying among business people is somewhat frank: “You should
not ask the frogs if you plan to drain their lake.” On the other hand, given the experi-
ence of the workshop, we must ask a more important question: Is the interlanguage
approach a concept of moderate change, or could it also be used for paradigm shifts
in business? Collins, Evans, and Gorman suggest a trajectory in figure 2.2, and it would
be worth exploring the conditions under which such a trajectory could be managed.
(b) Is the trading zone an evolutionary concept? It seems that time is not an issue in
a trading zone, whereas in business the windows of time are very narrow. A trading
zone must be fed and nurtured, so it is evolutionary by nature. Of course, trading over
a long period of time changes the traders’ minds. That is to be expected. It would not
be until long after the trading has become institutionalized that the zone itself could
become a new field, be it in science or technology. We may need to know more about
the time aspect of trading zones.
For a trading zone to react quickly, it needs to have an already established common
language. Consider the way in which American banks and the U.S. Federal Reserve
have worked together to bail out the American financial system. Surprising deals like
the Fed’s takeover of Bear Stearns came together very quickly, but the parties pretty
much spoke the same language, although they often operated from different
assumptions.
(c) Trading zones are not about homogenization. They maintain the distinct identities
of the parties involved but develop a common language. Classical examples are biology
and chemistry evolving into biochemistry; or chemistry, atomic physics, and engineer-
ing evolving into nanotechnologies. This sounds like the famous “merger of equals”
in business, which has failed so many times because the social interactions of large
organizations involve power, money, and emotions, all of which are brutal in their
influence on rational thinking. A trading zone is not an intrusion, nor is it coercive.
Instead, it is a structured site for peaceful exchange. First and foremost, it is an
exchange issue, not a turf issue. Perhaps integrating sciences is, by nature, a more
“rational” process than integrating businesses?
(d) Communication is at the core of the trading zone. Communication assumes ratio-
nal argumentation. However, we all know from business that many decisions are not
made or are made improperly due to emotional hurdles that cannot be overcome.
Many great concepts in business could not be agreed upon because the parties involved
were not ready to change their beliefs, taboos, and emotions.
Do trading zones deal with the nonrational part of changing people’s opinions?
Considering the question from the standpoint of business strategy, it seems there is
sufficient evidence that trading zones help us to improve our understanding of change
processes. Further exploration is necessary to reach the specifics of the nature of
boundary crossings in the context of big changes and small changes alike.
Boundary crossings have different requirements, depending on what has to be
crossed:
What are different crossing situations?

Is there a taxonomy of crossings?
What are the requirements for a crossing?
What are the consequences of crossing?
Are there different ways of crossing?
What are the typical crossing processes?
What are typical structures of crossing?
What is a successful crossing culture?
What is a “business creole”?
A framework of trading zones would be very helpful in business. As the workshop

has shown, different disciplines could contribute, and the chances are great that the
building blocks of the sociology of science could help us explore the business changes
in more depth. This process could prove to be a trading zone between the sociology
of sciences and business administration.
References
Collins, Harry, Robert Evans, and Michael E. Gorman. 2007. Trading Zones and Interactional
Expertise. Studies in History and Philosophy of Science 39 (1):657–666.
Chicago Press.
Gorman, Michael E. 2005. Trading Zones, Interactional Expertise and Interdisciplinary Collabora-
tion. Draft description of a workshop to be held 21–25 May 2005 at Arizona State University.
http://bart.tcc.virginia.edu/Tradzoneworkshop/Tradingzonesdesc.pdf, 1.
Malinowski, Bronislaw. 1922. Argonauts of the Western Pacific: An Account of Native Enterprise and
Adventure in the Archipelagoes of Melanesian New Guinea. London: Routledge.
Ratzel, Friedrich. 1899 [1891]. Anthropogeographie. 2d ed. 2 vols. Stuttgart: J. Engelhorn.
Schumpeter, Joseph A. 1975 [1942]. Capitalism, Socialism and Democracy. New York: Harper and
Row.
Simmel, Georg. 1908. Soziologie. Untersuchungen über die Formen der Vergesellschaftung. Leipzig:
Duncker & Humbolt.
III Ethics and Trading Zones
In this part, three chapters are devoted to the topic of ethics, values, and trading zones.
Michael E. Gorman and Patricia H. Werhane demonstrate how trading zones and
interactional expertise can be used to avoid normalized deviance in organizations.
Normalized deviance occurs when evidence that suggests a problem within an orga-
nization is reclassified as normal—and then future instances maintain that classifica-
tion, instead of serving as warning signs. Gorman and Werhane discuss two cases: the
explosion of the space shuttle Columbia and the collapse of WorldCom. In the former
case, a tiger team should have been assembled to decide what to do about the piece
of foam that struck the shuttle’s wing; instead, the foam strike was handled through
normal, compartmentalized bureaucratic channels, and the astronauts were told it was
not significant. In the case of WorldCom, an earnings shortfall was covered up by
adjustments to liability and expense accounts, on the grounds that the shortfall was
temporary. Instead, the problem grew larger each quarter, and more “creative account-
ing” occurred until WorldCom had to admit it had lost $11 billion during a period
when it had been declaring positive earnings. There were people both at NASA and
at WorldCom who saw that the “normalized” practices were deviant. A proper trading
zone within the organizations could have allowed this dissent to be aired. The larger
point is that the right kind of trading zone can facilitate the exercise of moral imagi-
nation, in which participants not only exchange goods and services, but also try to
understand each other’s mental models of a situation and seriously consider alterna-
tives that might otherwise have been dismissed.
Ann E. Mills, Mary V. Rorty, Lynn Isabella, and Donna T. Chen apply the trading
zones model to ethics and values in global health care. One of their examples is the
President’s Emergency Plan for AIDS Relief (PEPFAR), begun in 2003. They question
whether the primary goal of this plan was to prevent AIDS, or to promote values like
abstinence and the elimination of prostitution. Trading zones have to include consid-
eration of values from the beginning, and have to revisit values throughout the process
244 Ethics and Trading Zones
of adaptively managing initiatives that include multiple stakeholders, as PEPFAR does.

An example of such adaptive management is the development of the drug Gleevec®,
which showed promise of treating a type of leukemia during its early clinical trials.
The trading zone around this drug shifted its goal from proving safety and efficacy to
getting it to patients as quickly as possible, which meant working for FDA approval.
This kind of shift in goals and procedures based both on empirical results and values
considerations is characteristic of adaptive management.
Mary Ann Leeper, Elizabeth Powell, and Patricia H. Werhane use the example of
the Female Health Company (FHC) to show how a shift in goals transforms a trading
zone. The initial objective of the FHC was to develop and market a female condom
to over-the-counter customers. The market in the United States was limited because
there were many other options available to women, but a great potential market
existed in sub-Saharan Africa, where women were being devastated by AIDS. The
problem was that these women could not pay for the product. Therefore, the FHC had
to shift its goal from marketing a for-profit birth control product to building a trading
zone among organizations for HIV/AIDS prevention. This led to forming a new
model—an alliance. All the partners in this alliance had to recognize the urgency of
reducing or eliminating the risk of AIDS and had to accept that as their common goal.
Villagers and other stakeholders also had to be recruited. The end result is that the
FHC now operates in over a hundred countries using the alliance model, with its
primary goal the reduction of HIV/AIDS.
12 Using Trading Zones to Prevent Normalized Deviance in
Organizations
Michael E. Gorman and Patricia H. Werhane
Peter Galison has developed the notion of a “trading zone” to describe how people
from vastly different theoretical, practical, or cultural perspectives can interact mean-
ingfully about subjects which they understand from seemingly incommensurable
points of view (Galison 1997). A trading zone is a locus of communication, involving
the development first of a jointly understood jargon, then a pidgin, and finally a creole
among individuals or groups of individuals whose background or theoretical points
of view are vastly different and sometimes conflicting, such that no individual could
hold all represented points of view simultaneously without contradiction. Galison
defines a jargon as a set of shared meanings for terms. A pidgin is a language sufficient
for relatively simple trades constructed from elements of at least two languages; and
a creole is a pidgin extended to the point where it can serve as a new language. These
three levels of linguistic communication are on a continuum; a trading zone typically
begins with a few common terms, or jargon, quickly develops phrases that can be
learned quickly and understood across the zone, then eventually can lead to a new
language, or creole, that is taught to future generations.
Collins, Evans, and Gorman (2007, and chapter 2 in this volume) have pointed out
that coordination across a trading zone can be accomplished by interactional experts
playing a role similar to that of trade agents. The interactional expert is fluent in the
language of another disciplinary or stakeholder community, to the point that she or
he can pass as a member of that community—except when it comes to actually per-
forming the expertise. The interactional expert can play a role similar to a trade agent
in a trading zone, facilitating exchanges even when the multiple parties have not yet
evolved a creole.
The idea of a trading zone has normative dimensions. Trading zones are
useful whenever communication occurs across apparently incommensurable perspec-
tives, and, we shall argue, the engagement in a trading zone, coupled with moral
imagination, can result in rich, value-laden outcomes. When trading zones involve
246 Michael E. Gorman and Patricia H. Werhane
ethical conflicts, interactional expertise should involve a developed moral imagina-

tion. Here the interactional expert not only will have to master the language of
another disciplinary community, but will have to understand its values.
In particular, trading zones and interactional expertise can help organizations to
avoid the problem of normalized deviance. Normalized deviance occurs when indi-
viduals within organizations, often skilled professionals, playing what they believe to
be their well-defined professional roles, reclassify danger signals as within the bounds
of normality—and continue a course toward disaster. In the course of this chapter, we
will show how the formation of trading zones might have prevented the kind of nor-
malized deviance that contributed to the Columbia shuttle disaster and the collapse
of WorldCom.
Experts, Paradigms, and Particularized Worldviews
Expertise is clearly an essential part of a successful trading zone, but it can also make
communication across a zone difficult, because different expert communities operate
within unique paradigms. As Vaughan writes:
Implied in the term “expert” is some technical skill, gained either by experience, by profes-
sional training, or by both, that differentiates the professional from the lay assessment of risk.
Also implied is that professionalism will somehow result in a more “objective” assessment than
that of the amateur. But professional training is not a control against the imposition of par-
ticularistic worldviews on the interpretation of information. To the contrary, the consequence
of professional training and experience is itself a particularistic worldview comprising certain
assumptions, expectations, and experiences that become integrated with the person’s sense of
the world. The truth is that highly trained individuals, their scientific and bureaucratic proce-
dures giving them false confidence in their own objectivity, can have their interpretations of
information framed in subtle, powerful, and often unacknowledged ways. (Vaughan 1996,
63–64)
Vaughan’s description of the “particularistic worldview comprising certain assump-

tions, expectations, and experiences that become integrated with the person’s sense
of the world” corresponds to Thomas Kuhn’s notion of a paradigm. Kuhn (1970)
referred to activities carried on within a paradigm as normal science, a kind of standard
operating framework that is taken for granted by most experts within a domain. A
paradigm includes shared assumptions, generalizations, and examples that illustrate
the proper approach and methods. Large parts of a paradigm are tacit, which means
that to those within the specific expertise community, the paradigm becomes a kind
of “standard operating reality.”
Using Trading Zones to Prevent Normalized Deviance in Organizations 247
According to Kuhn, paradigms change when sufficient anomalies are encountered

—when results do not fit expectations often enough and significantly enough that
the paradigm itself becomes subject to question. One way of protecting a paradigm
is to engage in normalized deviance, which Vaughan (1996) defines as carrying on
as if nothing were wrong despite accumulating evidence of problems. Normalized
deviance is the management equivalent of ignoring anomalies that suggest the para-
digm no longer works. An anomaly in science can, for example, be classified as an
error, perhaps because of faulty procedures, or chance, or a variety of other legitimate
concerns when one is performing a complex experiment or making an extremely
difficult observation (Gorman 1992). In the literature on the psychology of science,
this phenomenon is described as “confirmation bias”—the tendency to seek evidence
that supports a hypothesis, or to reinterpret potential negative evidence as an error.
Reclassifying as errors results that appear to deviate from the dominant hypothesis
allows them to be “normalized.” All future events of this class can then be labeled
errors, and this quick recognition of supposed errors can become part of one’s
expertise.
Normalized deviance can have devastating consequences in engineering and busi-
ness. Vaughan studied the case of the failure of the O rings during the launch of
the Challenger space shuttle in 1986. Although this potential failure had been high-
lighted before the launch by engineers like Roger Boisjoly, the O rings’ partial erosion
was reclassified as an acceptable risk. The O rings that appeared to have been com-
promised were classified as errors due to factors like putty defects, hair on the ring,
or cold temperatures. This last factor was the critical one in the O ring failure, but
it was not a consistent signal, for O rings had also been compromised on flights that
took place in temperatures above 70 degrees. There were multiple ways of explaining
partial erosion of an O ring that did not undermine the consensus that such erosion
was an acceptable risk, especially because each O ring was backed up by another,
redundant one.
Vaughan (1996, 249) quotes Larry Wear, Marshall Space Flight Center’s manager
of solid rocket motors, as saying: “Once you’ve accepted an anomaly … you’re at the
point that it’s very hard to draw the line. You know, next time they say it’s the same
problem, it’s just eroded 5 mils more.” Here is the quintessential statement about
normalized deviance: what was once an anomaly is now an acceptable risk, and is
hard to reclassify as a danger, even when there are signs that the risk has become
greater. In the case of the Challenger disaster, successful flight after successful flight
appeared to confirm that the risk was acceptable.
Mental Models
To better understand the process of normalized deviance, we will need to push beyond
Kuhn’s provocative but vague idea of a paradigm to consider two more precise con-
cepts from the cognitive science literature: mental models and scripts. As defined by
Peter Senge (1990, 8), “Mental models are deeply ingrained assumptions, generaliza-
tions, or even pictures or images that influence how we understand the world and
how we take action. Very often, we are not consciously aware of our mental models
or the effects they have on our behavior.”
Mental models might be hypothetical constructs of the experience in question or
of scientific theories; they might be schemas that frame the experience, through which
individuals process information, conduct experiments, and formulate theories. Mental
models function as selective mechanisms and filters for dealing with experience. In
focusing, framing, organizing, and ordering what we experience, mental models can
bracket and leave out data, and emotional and motivational foci can taint or color
experience. Creative individuals can come up with new mental models that allow
them to frame a situation differently, leading, among other things, to scientific dis-
coveries and technological innovations (Gorman, 1998). Each of us can frame any
situation, event, or phenomenon in more than one way, and that same phenomenon
can also be socially constructed in a variety of ways.
There are “different moral consequences depending on the way we frame the
situation” (Johnson 1993). Our views of the world, of ourselves, of our culture and
traditions, and even our values orientation are all constructions—all experiences
are framed, ordered, and organized from particular points of view. These points of
view or mental models can be incomplete, sometimes distorted, narrow, and single-
framed, or they can be creative and dynamic. Because they are learned, mental
models can be modified, though it can be difficult to change the way a situation
is framed.
According to this point of view, a paradigm is a “deeply ingrained” mental model
that influences both understanding and action. As a mental model, a paradigm has
both explicit and tacit components: some aspects of the paradigm can be articulated
by those in it, while other aspects unconsciously structure thought. Consider Kuhn’s
account of what it is like for experts in a community after the shift to a new
paradigm:
Led by a new paradigm, scientists adapt new instruments and look in new places. Even more
important, during revolutions scientists see new and different things when looking with familiar
instruments in places they have looked before. It is rather as if the professional community had
been suddenly transported to another planet where familiar objects are seen in a different light
and are joined by unfamiliar ones as well. (Kuhn 1970, 111)
To put this in the language of cognitive science, during a paradigm shift, scientists
experience a radical shift in a fundamental mental model that changes the way they
look at problems across their domain. A good example is Kepler’s abandonment of
one of the accepted realities of his day: that the planets had to move in perfectly
circular orbits because they were mounted on spheres. Kepler, using the best data and
mathematics of his day, could not fit the orbit of Mars into a perfect circle. Instead
of accepting the perfect circles as reality, he was able to abandon this mental model
and fit the orbit of Mars to an ellipse (Gorman 1998). In fact, it turns out that all
planetary orbits are elliptical.
Mental models can operate in design as well as science. Alexander Graham Bell’s
invention of the telephone depended on a mental model based on the human ear,
whose function he understood intimately (Gorman 1997). For Bell, this mental model
suggested how to frame the problem of transmitting and receiving speech. He built a
kind of simple electromechanical ear, with speaking tube, diaphragm, an armature
that worked like the bones of the ear, and finally an induction coil to turn the motions
into current.
One of Bell’s rivals, Elisha Gray, also thought about using the ear as a model for a
telephone, indicating that Bell’s insight about the ear was not a new paradigm—it did
not move research in this area to fundamentally new ground. What Gray and other
telegraph inventors lacked was Bell’s intimate understanding of the ear and how it
translated the motion of the eardrum into a series of undulating waves. Bell wrote a
patent covering all devices that transformed sounds into an undulating electric current,
thereby claiming an enormous territory that no other telegraph inventor would have
considered worthwhile (Gorman 1997).
Mental models account not only for the advantages but also for the limitations of
a new approach or an existing one. The “ear” mental model is both the strength and
weakness of Bell’s approach: it was great for building a receiver, but a poor design for
a transmitter, because the ear itself functions only as a receiver.
Scripts
How do mental models get translated into the kind of standard operating procedure
that allowed the O rings to continue to be regarded as an acceptable risk? Consider
the case of the Ford Pinto, a compact car that was designed in record time with a
“limits of 2,000” constraint: it could exceed neither 2,000 pounds in weight nor $2,000
in cost.1 Because of the speed with which it was designed, a potential problem with
the fuel tank was identified only after the design was frozen: the tank ruptured when
hit from the rear at a relatively low speed (about 30 mph).
The Pinto was introduced in 1970. Dennis Gioia was a relatively new employee at
Ford in charge of recalling defective automobiles. When provided with evidence that
a number of Pintos had exploded, including “graphic, detailed photos of the remains
of a burned-out Pinto in which several people had died” (Gioia 1992, 382), he did not
issue a recall. Looking back on his decision, Gioia concluded, “My own schematized
knowledge influenced me to perceive recall issues in terms of the prevailing decision
environment and to unconsciously overlook key features of the Pinto case, mainly
because they did not fit an existing script. Although the outcomes of the [Pinto] case
carry retrospectively obvious ethical overtones, the schemas driving my perception
and actions precluded consideration of the issues in ethical terms because the scripts
did not include ethical dimensions” (Gioia 1992, 385).
“Schema” is a term some cognitive scientists use to refer to the expectations a
problem solver brings to a situation. Kuhn’s paradigm is a kind of higher-level schema
that tells scientists where to look for interesting research problems, how to investigate
them, and what results to expect. A mental model functions like a schema because it
embodies expectations, for example assumptions about the motions of planets or what
will happen if an armature is built like the bones of the human ear.
Gioia’s set of expectations led him to look for a high frequency of incidents or a
clear cause. At the time, he was not aware of the test analysis that showed a problem
with the fuel tank design, so instead he looked at the accidents and found no pattern—
just occasional flaming crashes. Part of Gioia’s schema was the assumption that all
small cars are more prone to serious crashes of this sort and that people accept risks
when they drive—because “safety doesn’t sell.” Gioia was also experiencing cognitive
overload—he had lots of cases and reports to attend to. So the crashes did not consti-
tute an anomaly for him.
Schemata and mental models can be translated into scripts that tell us how we are
supposed to behave in certain situations (Schank and Abelson 1977). As an example,
most of us have scripts for restaurants that include waiting to be seated, getting a
menu, ordering, and paying a bill. After seeing one particularly gruesome wreck, Gioia
activated one of his standard scripts by calling for a preliminary review of the Pinto
case. Similar reviews occurred repeatedly and at multiple levels in the Challenger case.
These reviews are scripted in the sense that the participants know their respective roles
and rarely step out of them. The group convened by Gioia unanimously decided to
leave the Pinto on the market.
Challenger, like other space shuttles, was funded on the assumption that it would
be a safe and cost-effective way to put satellites, telescopes, and other equipment into
space. It was believed to be no more dangerous than flying a 727 to Disney World.2
The night before the Challenger launch, a group of engineers at Morton Thiokol called
a special review session with NASA engineers at the Marshall Space Flight Center to
decide if the launch should be postponed because it was going to be unusually cold.
Roger Boisjoly, an expert on O rings, was particularly concerned that there was no
data on O ring performance at temperatures below 53 degrees Fahrenheit. Marshall
Space Flight Center’s Lawrence Mulloy responded, “My God, Thiokol, when do you
want me to launch, next April?” (Vaughan 1996, 6). Marshall managers objected to
the equivocal nature of the data: the O rings had failed at both low and high tempera-
tures. The launch review script called for a lot of pushback from the managers, but
also an understanding that if Thiokol recommended the launch be postponed, NASA
would comply. Marshall managers asked the engineers if they could prove the O rings
would fail. Boisjoly’s data suggested that the probability of O ring failure increased
with lower temperature, but the relationship was less than perfect. The Thiokol engi-
neers went off-line to consult among themselves. Robert Lund, the vice president of
engineering, was asked by his immediate superior to “take off your engineering hat
and put on your management hat” (Werhane 1991, 606). He switched hats, the launch
was approved, and all seven of the crew perished.
When Lund switched from thinking as an engineer to thinking as a manager, he
was switching roles in the script, and this tipped the balance. Similarly, Gioia was
playing his role as a conscientious Ford employee, looking for hard evidence in the
design review that the original decision to launch the Pinto was wrong. Here is the
root of “normalized deviance,” in scripts and roles carried out within mental models
like “the shuttle is as safe as a 727” and “everyone knows compact cars are more
dangerous.”
As a result of his experience, Gioia (1992, 388) calls for integrating ethical decision
making into scripts through the use of “vicarious or personal experiences that inter-
rupt tacit knowledge of ‘appropriate’ action so that script revision can be initiated.”
This kind of vicarious or personal experience that interrupts an ongoing script neces-
sarily involves moral imagination.
Moral Imagination
Moral imagination is the extension of mental models into ethics. “Moral imagination”
includes “the awareness of various dimensions of a particular context as well as the
operative framework and narratives. Moral imagination entails the ability to under-
stand that context or set of activities from a number of different perspectives, the
actualizing of new possibilities that are not context-dependent, and the instigation of
the process of evaluating those possibilities from a moral point of view” (Werhane
1999, 5).
Moral imagination involves:
1. Becoming aware of one’s own mental model and scripts in a particular situation.
Trading zones can be a first step in stimulating this kind of awareness, because
exchanges in these zones occur among parties with very different mental models.
2. Imagining different mental models from which the situation could be viewed. Here
is where Gioia’s (1992, 388) “vicarious or personal experiences that interrupt tacit
knowledge of ‘appropriate’ action” becomes particularly important—these experiences
suggest alternate mental models and scripts. One way to get this kind of experience
is to consult others—particularly those who can take an outsider’s view, who are not
part of the “normalized deviance” that takes place in some organizations. As Senge
writes, “Each person’s view is a unique perspective on a larger reality. If I can ‘look
out’ through your view and you through mine, we will each see something we might
not have seen alone” (Senge 1990, 248).
3. Evaluating the situation from the standpoint of both old and new mental models,
and envisioning alternate courses of action.
4. Coming up with a better mental model of the situation and more options for
resolving it.
Trading zones create the opportunity for exposure to different mental models of a
system and different scripts for dealing with situations. Therefore, trading zones can
spark moral imagination, by showing a participant in a zone that there is a very dif-
ferent way of viewing a situation, and thus a different way of behaving, when faced
with a problem.
Superordinate Goals
One way of sparking moral imagination in a trading zone is to have participants

working urgently toward a common goal, as when theoretical physicists and radio
engineers worked together to develop radar at the beginning of World War II. The
social psychologist Muzafer Sherif (1976) calls such a goal a “superordinate goal.”
Growing up in Turkey during the decline of the Ottoman Empire, Sherif saw firsthand
how members of groups would exhibit compassion toward their own and hostility
toward members of other groups. He decided to devote a career to understanding this

phenomenon (Gorman 1981). He and his wife, Carolyn, created a summer camp for
boys in Oklahoma at which each boy was randomly assigned to one of two groups,
and competition between the groups was encouraged (Sherif 1976). The Sherifs were
surprised at how quickly intergroup hostility developed.
They found the best way to bring the groups together was to introduce a superor-
dinate goal. Sherif fabricated a problem with the camp’s water supply, one that the
two groups had to work together to solve. Next, the groups had to work together to
get their lunches after an active morning. These kinds of common challenges were
successful at ending hostilities between the two groups. Sherif thought a superordinate
goal was superior to working together against a common enemy because having an
enemy perpetuates the problem of intergroup hostility.
Sherif’s campers were from homogeneous backgrounds, and were randomly sepa-
rated into their groups. They therefore did not mimic the long-term divisions between
religious, ethnic, and income groups that make intergroup hostility difficult to over-
come. Even without such divisions, moral imagination can be difficult to achieve. For
example, although access to food and water is a chronic problem for millions of
human beings, those who have abundant supplies do not see ending starvation as a
superordinate goal. They are missing the moral imagination necessary to see another’s
suffering as if it were one’s own—to see that ensuring that no child dies of thirst, or
starvation, is a superordinate goal for all.
Trading zones that incorporate these kinds of superordinate goals will require
participants to exercise moral imagination. But the participants will have to agree on
the superordinate goal, and see the urgency of achieving it. Let us consider two case
studies.
The Columbia Case
The foam did it, the institution allowed it.

—Columbia Accident Investigation Board member Duane Deal3
What was the superordinate goal of the shuttle program? A safe and efficient method
for putting satellites into space? Or a high-risk adventure that pushed the limits
of technology? It turns out the latter was more descriptive of the shuttle program
than the former, more explicit goal. Nevertheless, multiple stakeholders were told
the program was safe, including President Reagan, who said after the fourth shuttle
mission, “beginning with the next flight[,] the Columbia and her sister ships will be
fully operational, ready to provide economical and routine access to space for scientific
exploration, commercial ventures, and for tasks related to the national security.”4
The Challenger disaster of 1986 should have been a significant anomaly for those
whose mental model of the shuttle was a safe, reliable cargo carrier. But the O rings
were not the only recurring problem with the space shuttles. Beginning with the first
space shuttle flight, each shuttle had shed debris from the foam tiles that encased
much of the orbiter’s surface, and these tiles had caused at least some damage to every
flight. Originally, such damage was considered unacceptable, but as shuttle after
shuttle flew despite foam debris, what was once seen as deviant became normal. Foam
strikes went from being classified as “in-flight anomalies” requiring resolution before
the next flight to being “acceptable risks.” A particularly large piece of foam detached
from the shuttle Atlantis’s left bipod ramp in 2002, but even this did not push tile
debris from an acceptable risk to an anomaly.
The Columbia launched on January 16, 2003; eighty-two seconds into the launch,
a chunk of insulating foam detached from a fuel tank and struck the left wing. This
incident was not discovered until the morning of January 17, during a review of pho-
tographs of the launch—and because the photograph was blurry, it was hard to deter-
mine the size, shape, and momentum of the debris, and exactly where it had struck.
Budget cuts over the years had reduced the ability to keep better photographic records.
Analysts classified this event as “out of family,” triggering a script that called for
assembling a tiger team to investigate it. Instead, NASA created a Debris Assessment
Team (DAT) that reported to the Mission Management Team (MMT) headed by Linda
Ham. Rodney Rocha, a NASA engineer who was cochair of the DAT, asked at one
point if one of the astronauts could do a space walk to assess the damage, but he
never received an answer (Bohmer, Edmondson, and Roberto 2004). Rocha and the
DAT also decided to ask that Defense Department satellites try to capture images of
the wing. When Ham and the MMT found out about this request, which had not
been made through the usual bureaucratic channels, they canceled it. She did not
check to see if the DAT had asked for it. Rocha wrote an angry e-mail protesting this
decision—but sent it only to his colleagues, not to management. Later he explained
that “engineers were often told not to send messages much higher than their own
rung in the ladder” (Bohmer, Edmondson, and Roberto 2004). According to Vaughan,
“The MMT was supposed to operate in a decentralized way. That’s not what happened.
It was a very centralized operation with all the information flowing to the manager.
It was very rule-oriented, very protocol-oriented” (quoted in Bohmer, Edmondson,
and Roberto 2004). The MMT’s script thus took precedence over the essentially script-
less DAT.
Ham and the DAT operated from an overall mental model that indicated foam
debris was not a significant risk. They also thought there was little that could be done
even if imagery showed there was damage to the shuttle wing. The Columbia Accident
Investigation Board, convened after the tragedy, disagreed, but only in a qualified way.
They argued it might have been possible to launch a rescue shuttle—but of course,
that mission might have experienced a foam strike as well.
The Columbia crew was informed of the debris strike in a call to Commander Rick
Husbands, who was assured it represented no threat. But the Columbia exploded upon
reentry because of wing damage caused by the debris; and all its astronauts
perished.
Moral imagination certainly would have helped in this case: for example, decisions
could have been influenced by imagining how each decision would look to someone
outside of the culture, outside of NASA—say a member of one of the astronauts’ fami-
lies. This kind of shift of perspective can prevent normalized deviance, but only if the
organization encourages communications across hierarchical boundaries, such as
those created by NASA. According to Richard Feynman, who served on the Presidential
Commission evaluating the Challenger disaster, “It appears that there are enormous
differences of opinion as to the probability of failure with loss of vehicle and human
life. The estimates range from 1 in 100 to 1 in 100,000. The higher figure comes from
working engineers, and the very low figures from management” (quoted in Vaughan
1996, 274). Feynman himself felt the probability was about 1 in 25; as the Columbia
disaster illustrated (unfortunately), his estimate was close to the mark.
Here is where trading zones could help. In hierarchical organizations, trading zones
are the informal exchanges of favors and information that allow communication and
collaboration horizontally, without going through the chain of command. These
trading zones can help different parts of the organization begin to see an overall
pathology, and get the support from each other to deal with it. In the case of the
space shuttle disasters, such a trading zone would have allowed working engineers
who thought the probability of failure was one mission in a hundred to interact not
only with managers who thought the chance was much more remote but also with
policymakers who assumed the system was safe. During the development of radar, for
example, the military often tried to get the MIT RadLab to simply develop devices
according to their specs. RadLab responded by telling Navy officers to “bring back
your man who understands radar, you bring your man who understands the Navy,
who understands aircraft, you bring your man who understands tactics, then we’ll
talk about your needs” (quoted in Conant 2002, 256). RadLab transformed the mili-
tary’s top-down decision model into a trading zone.
After the Columbia foam strike occurred, and while the shuttle was still in orbit, a
tiger team could have been made up of people who had good “tribal knowledge” of
both where to find appropriate experts at NASA and how to navigate the bureaucracy
to mobilize them. At least some team members would have had to possess sufficient
interactional expertise to appreciate insights from those with different training and
backgrounds. This tiger team would have to include managers responsible for the
launch decision, who would still have to make fast decisions but would know they
could get the best knowledge in the organization mobilized to help them. Such a tiger
team would also have been in touch with the astronauts themselves, sharing knowl-
edge and involving them in decisions. This kind of trading zone was set up by the
Kennedy administration during the Cuban missile crisis, when officials and experts
from multiple levels of the administration met and considered alternative solutions;
the president listened and participated but did not dominate (Janis 1982).
A proper tiger team trading zone that included Rocha, Ham, the astronauts, and
others in an equal exchange would at least have led to the acquisition of more infor-
mation that might have revealed the anomalous and dangerous nature of the foam
strike. But the best use of trading zones is to avoid, as much as possible, getting stuck
in such crisis situations. Both moral imagination and trading zones depend, in this
case, on agreement on a superordinate goal, and here is the deepest difficulty. The
MMT at NASA was certainly concerned with safety, but in a context that emphasized
the quick turnaround of shuttles that might still develop into reliable cargo haulers
to the space station and for defense satellites. But Columbia was being used to do
science in space, which suggests a more appropriate goal—the shuttle program as a
technoscientific experiment. This model or orientation would have removed the time
pressure to fly a lot of missions and instead would have put a premium on learning
from each flight. Irritating problems like the O rings and the foam debris would have
then turned into opportunities to improve future missions. Safety would still have
been at a premium, but everyone would have understood the risks involved and
expected, alas, that at some point some astronauts would die.
One trading zone at NASA could have been centered on the values of safety and
saving astronauts’ lives. Everyone at NASA thought safety was of paramount impor-
tance, but there was no informal network that regularly exchanged opinions about
safety across the organization. Such a zone should have included participants repre-
senting different roles and areas of expertise, like astronauts, engineers from NASA,
engineers from subcontractors like Thiokol, management, and those policymakers
who funded the mission. This kind of a trading zone would have been a good place
to continually revisit the tradeoffs between safety and mission goals. From the stand-
point of the different stakeholders, what gains justify the risks? How can these risks
be minimized? How should different stakeholder groups be informed about anomalies
that might represent new risks?
The Columbia Accident Investigation Board made a recommendation that is rele-
vant to the problem of normalized deviance: that NASA should create an independent
Technical Engineering Authority which would, among other things, decide what
constituted an anomalous event.5 Unfortunately, this recommendation does not
address the root of the problem. Instead of creating a trading zone among different
areas of expertise, it mandates that one group of experts have authority; but, as we
have seen, if their mental model does not line up with all the various complex dimen-
sions of the situation, they may fail to recognize anomalies. Bringing people from
multiple areas of expertise together does not guarantee that anomalies will be recog-
nized, but it increases the probability, because different mental models and scripts
would be applied to the situation. An equal trading zone with a creole ensures that
these different perspectives will be exchanged and heard. If the trading zone includes
management, as it should, then the recommendations are likely to be implemented.
The overall process of using trading zones is iterative: the results of management
strategies that emerge from trading zones have to be evaluated in trading zones. A
complex system like the shuttle will always produce new data that might represent a
significant anomaly or that could legitimately be classified as confirmation of the
existing management approach. The only way to be sure is to continue the dialog
across the trading zone. New stakeholders may be added who learn the creole and/or
gain interactional expertise, and old ones may depart, but the exchange has to con-
tinue. This did not happen after the Challenger disaster, so it should not be surprising
that the Columbia disaster occurred. What is most tragic is that the second failure may
have been avoidable.
The WorldCom Case
WorldCom put extraordinary pressure on itself to meet the expectations of securities analysts.
This pressure created an environment in which reporting numbers that met these expectations,
no matter how these numbers were derived, apparently became more important than actual
financial reporting.
—Former Attorney General Richard Thornburgh, in Jeter 2003, 194
Normalized deviance can occur in almost any type of organization or culture, wherever
evidence of mounting problems is reclassified as acceptable risk or even normal prac-
tice. On the other hand, trading zones created around superordinate goals require
collaboration among those whose perspective and areas of expertise are different,
increasing the probability that anomalies will be recognized and new solutions evolved.
The rise and fall of WorldCom illustrates how reclassifying mounting problems as
acceptable risks can lead an organization to ruin.6 WorldCom began in 1983 as Long
Distance Discount Services (LDDS), a small Arkansas-based broker of long-distance
telephone service. After the court-ordered breakup of AT&T’s Bell System and subse-
quent deregulation of the long-distance market, LDDS—like many other reseller
companies—began buying large blocks of capacity from the larger long-distance com-
panies at reduced rates and reselling that time to small business and individuals at a
slight markup over its cost. In the beginning, the company was not able to earn
enough to cover its expenses.
Bernie Ebbers, one of the founding investors, took over the management of LDDS.
Ebbers directed a continuing stream of acquisitions in the mid 1990s, expanding
the scope of the company’s operations and renaming the company WorldCom. In
1995 WorldCom reported revenues of $3.9 billion, operating out of 160 offices
worldwide, with 7,000 employees. With its acquisition of MFS Communications
Company in 1996, WorldCom became the owner of UUNet, a key force in Internet
access. Then, in 1998, WorldCom stunned the business world with its purchase of
MCI, the nation’s second-largest long-distance carrier and triple its size. In 1999,
WorldCom and Sprint agreed to merge, but U.S. and European regulators stopped
the merger in 2000. With no other big telecommunications companies to acquire
or with which to merge, Ebber ended his strategy of growth through acquisition.
Management had never taken the time to integrate the systems of its multiple
acquisitions: for example, there were at least sixty billing platforms. Some customers
were visited by two different WorldCom representatives, so that WorldCom was
often competing with itself.
One strategy for dealing with this mix of different cultures would have been to
establish trading zones across the different companies, linking experts in areas like
billing, accounting, technology, and customer relations in a network that was encour-
aged, but not dominated by, management. Such a zone could have led to an intra-
company creole that spanned the different acquisitions. An important motivation for
such a zone would be a shared sense of mission, or superordinate goal. In this case,
the goal might have even emerged out of the trading zones, as participants came to
respect each other and recognize they could do things together that none could do
only within her or his heritage organization.
Most of WorldCom’s employees had no idea whether the company was profitable,
in large part because its accounting methods were described as “back pocket.” Con-
cerns began to crop up. That internal confusion could not have come at a worse time,
as the Internet bubble was collapsing and many of WorldCom’s customers were forced
to declare bankruptcy. Even as the industry was slowing down, Wall Street maintained
high expectations for WorldCom. Failure to meet analysts’ expectations would acceler-
ate the decline in the stock price, with potentially serious consequences: most of
WorldCom’s senior people owned significant numbers of shares and at current prices
were very wealthy, at least on paper.
In October 2000, Scott Sullivan, WorldCom’s CFO, discovered that the cost of
leasing lines was much higher than expected, but it was not clear why (Cooper 2008).
Financial results would be made public in a few days, and the stock would be ham-
mered. Sullivan asked accounting executive Betty L. Vinson and her colleague Troy
Normand to make adjustments to liability and expense accounts to cover a shortfall
of $828 million. Vinson was upset by this request because there was nothing in the
business climate that would justify a reduction in reserves for possible bad debts—in
fact, given the circumstances, those reserves ought to be increased. Further, there was
no logic to reducing an allowance for bad debts and using that drawdown to reduce
line costs. She was assured that no one was happy that these entries needed to be
made and that it would never happen again.
In taking this step, Sullivan was reclassifying an apparent anomaly as an error. He
did not know what had caused the accounting shortfall, but he assumed it was not a
real signal of danger. Even if this shortfall turned out to be real, he assumed it could
be made up next quarter, because the telecommunications business had to get better
and WorldCom was in good shape to benefit from an upturn. Here Sullivan was
operating from a false mental model about the industry and its prospects.
Vinson and Normand made the entries as requested, but decided to resign. However,
when Sullivan heard about their impending resignations, he called them into his
office. He explained that he was trying to fix the company’s problems and urged them
to think of the company as an aircraft carrier: “We have planes in the air. Let’s get
the planes landed. Once they are landed, if you still want to leave, then leave. But
not while the planes are in the air.”7
Sullivan assured both Vinson and Normand that they had done nothing wrong,
and that he would assume all responsibility for the entries. Furthermore, he said he
would lower earnings guidance for the next quarter, to compensate. For Vinson, family
security was an important issue. She was the family’s chief breadwinner, earning more
than the $40,000 her husband brought home annually. In addition, the family
depended on her health benefits. It was clearly going to be very difficult to replicate
that salary in a small Mississippi town. She and Normand both wrote letters of
resignation, but decided they would not send them unless Sullivan failed to lower the
earnings projections. He did as he promised, they kept the letters, and they hoped the
problem was solved.
It was not. Earnings kept dropping, and Vinson and Normand had to do more
creative accounting. Sullivan instructed them to count excess line costs as capital assets
instead of expenses, because capital assets are written off over a longer time period,
buying more time for the hoped-for turnaround in the telecommunications
industry.
Here we see the signs of normalized deviance. Instead of looking at the unexpected
shortfall as a sign the company was not making money, the CFO rationalized that it
was a temporary anomaly which could be papered over by pushing the boundaries of
accounting rules. This was the first step down a slippery slope. When earnings con-
tinued to fall, more creative accounting seemed to be the only answer. The guiding
assumption was that problems would be short-term, despite the fact that the com-
pany’s growth-through-acquisition strategy was no longer feasible. But even if the
telecommunications business took off again and the company managed to properly
integrate its acquisitions—both unlikely scenarios—the accounting deceptions would
still have to be reported. The more effort the two accountants and the CFO spent
covering up, the worse the consequences would be when the house of cards fell.
A dose of moral imagination would have helped all three, but here we will focus
on Vinson’s dilemma. She could have imagined how the situation would look from
the perspective of another member of her profession outside of WorldCom. She knew
that what she was doing was not right—she wanted to resign, but got talked out of it
by an incorrect mental model of the situation, that the company was like an aircraft
carrier with planes in the air. If WorldCom was an aircraft carrier, it was sinking! She
was, in fact, likely to lose her job anyway, when the ship sank.
But Vinson also had a refuge in her codes of professional conduct. Trading zones
do not replace codes of conduct and mechanisms for reporting violations. For example,
WorldCom could have had a hot line to report unethical practices while preserving
confidentiality. Similarly, in the case of the Challenger, O ring engineers at Morton
Thiokol could have called the CEO of the merged company directly to report their
dilemma (Werhane 1991).
Perhaps if Vinson had been part of a cross-institutional trading zone, she would
have found the support to say no. One wonders, for example, why she did not contact
Cynthia Cooper, vice president of internal audit for WorldCom, who did see these
accounting procedures as deviant and eventually confronted Vinson to find out why
she had signed off on them—too late, after the damage was done.
Bernard Ebbers resigned as CEO of WorldCom on April 29, 2002. WorldCom COO
John Sidgmore succeeded Ebbers and immediately asked for a complete examination
and assessment of the books in every division of the company. Cynthia Cooper was
happy to comply because she had begun to suspect accounting irregularities, if not
outright fraud. In one disturbing conversation in March of that year, John Stupka,
head of WorldCom’s wireless division, had complained to Cooper that CFO Scott
Sullivan had yanked $400 million from Stupka’s division to use as an income boost
for the company. Stupka had specifically set aside that money to make up for potential
shortfalls; now his division would have to report a tremendous loss in the coming
quarter. He had approached two Arthur Andersen auditors, who backed Sullivan’s
decision. Stupka and Cooper were skeptical of that decision, however. They knew that,
under accounting rules, a company must set up a reserve to cover any debts not col-
lected in order to avoid reflecting too high a value on its books. Had Stupka been in
a trading zone with other managers in other parts of WorldCom facing similar pres-
sure, he and Cooper might have gotten a more complete picture of the problem and
found allies for immediate action.
The Securities and Exchange Commission (SEC) grew suspicious because, although
2001 was a notoriously difficult year for telecommunications companies, only World-
Com had a successful year financially. As a result, the SEC submitted a “Request for
Information,” giving the commission access to the company’s financial data unavail-
able through public sources. Startled by the request, Cooper and her staff began to
compile pertinent information, often working at night. She and her staff discovered
$2 billion in suspicious accounting entries.
Sullivan asked Cooper and her colleagues to hold off on the audit and not to reveal
its results, explaining that he planned to take care of the problems the following
quarter. But Cooper refused to stall. She talked to a number of employees, including
Betty Vinson, who admitted she made the entries Sullivan requested without being
able to properly justify them.
On June 25, 2002, WorldCom revised its financial statements over the previous five
quarters to reflect the accounting errors, which amounted to $3.8 billion. The stock
plunged to nine cents. As investigations continued, the accounting fraud was eventu-
ally found to have been $11 billion, the largest in corporate history. CFO Sullivan
testified against CEO Ebbers; the former was sentenced to five years in prison, the
latter to twenty-five.
The anomaly that threatened WorldCom’s business paradigm was the fact
that the company was losing money. “Deviant” information that revealed the
company’s problems was “normalized” by a variety of accounting tricks. Those
who participated in these frauds hoped that the whole telecommunications busi-
ness would eventually improve to the point where the company would make
money again—ignoring the fact that the company’s growth was based on acquisi-
tions that were never properly integrated. Trading zones across these different parts
of the organization certainly would have helped WorldCom, both from a business
and an ethical standpoint.
Conclusion
In the case of the Columbia disaster, the value of trading zones that provide alternative
mental models and scripts is clear. Trading zones cannot make a high-risk adventure
in space safe, but they can lead to a better understanding and acknowledgment of
risks and a consideration of more alternatives in a crisis situation.
In cases like WorldCom, where companies acquire or merge with other companies
that have very different cultures, trading zones have real value in creating informal
communication networks across the apparently incommensurable divisions in prac-
tices, attitudes, and loyalties. Such trading zones may not have made WorldCom
profitable, but they would at least have made it clear what the company’s actual state
was, and facilitated truth telling with investors and shareholders. Trading zones make
anomalies more salient, and evolve strategies for coping with them.
Notes
1. This account of schemas, scripts, and the Pinto case is adapted from Gorman 1998.
2. Former shuttle astronaut Dr. James Bagian recalled that, during this time, NASA often likened
flying the shuttle to flying a 727 plane to Disney World. However, this assertion quickly became
a running joke among astronauts at NASA, due to the difficulties of completing the projected
fifty missions in one year. “That was the party line,” explained Bagian. “If flying a shuttle was
like flying a 727 to Disney World, there wouldn’t be anybody going to Disney World!” (Bohmer,
Edmondson, and Roberto 2004, 5).
3. Quoted in Michael A. Roberto, Richard M. J. Bohmer, Amy C. Edmonson, and Erika M. Ferlins,
teaching note to Columbia’s Final Mission case, HBS no. 5-305-033, p. 7.
4. Quoted in ibid., 11.
5. The Columbia Accident Investigation Board report, vol. 1, August 2003, also included a long
list of technical recommendations and additional organizational ones. For all volumes of the
report and appendices, go to http://caib.nasa.gov/news/report/default.html.
6. Unless otherwise attributed, details of the WorldCom case are drawn from two Darden School
of Business case studies prepared by Jenny Mead under the direction of Patricia H. Werhane,
Ruffin Professor of Business Ethics: “Cynthia Cooper and WorldCom A,” UVA E-2079, and
“WorldCom: Keeping Planes in the Air” (in preparation). Go to www.darden.edu to obtain the
cases.
7. Susan Pulliam and Deborah Solomon, “How 3 Unlikely Sleuths Uncooked WorldCom’s Books;
Company’s Own Auditors Sniffed Out Cryptic Clues, Followed Their Hunches,” Wall Street Journal
Europe, October 31, 2002, A1.
References
Bohmer, R. M. J., A. C. Edmondson, and M. A. Roberto. 2004. Columbia’s Final Mission. Business
case no. 9-304-090. Harvard Business School, Cambridge, MA.
in History and Philosophy of Science 39 (1):657–666.
Conant, J. 2002. Tuxedo Park: A Wall Street Tycoon and the Secret Palace of Science That Changed
the Course of World War II. New York: Simon and Schuster.
Cooper, C. 2008. Extraordinary Circumstances. Hoboken, NJ: Wiley.
Chicago Press.
Gioia, D. 1992. Pinto Fires and Personal Ethics. Journal of Business Ethics 11:384–385.
Gorman, M. E. 1981. Pre-war Conformity Research in Social Psychology: The Approaches of Floyd
H. Allport and Muzafer Sherif. Journal of the History of the Behavioral Sciences 17:3–14.
Gorman, M. E. 1992. Simulating Science: Heuristics, Mental Models and Technoscientific Thinking.
Bloomington: Indiana University Press.
Gorman, M. E. 1997. Mind in the World: Cognition and Practice in the Invention of the Tele-
phone. Social Studies of Science 27 (4):583–624.
Gorman, M. E. 1998. Transforming Nature: Ethics, Invention and Design. Boston: Kluwer Academic
Publishers.
Janis, I. L. 1982. Groupthink: Psychological Studies of Policy Decisions and Fiascoes. 2nd ed. Boston:
Houghton Mifflin.
Jeter, L. W. 2003. Disconnected: Deceit and Betrayal at WorldCom. Hoboken, NJ: Wiley.
Kuhn, T. S. 1970. The Structure of Scientific Revolutions. Rev. ed. Chicago: University of Chicago
Press.
Schank, R., and R. Abelson. 1977. Scripts, Plans, Goals and Understanding: An Inquiry into Human
Knowledge Structures. Hillsdale, NJ: Lawrence Erlbaum Associates.
Senge, P. 1990. The Fifth Discipline: The Art and Practice of Learning Organizations. New York:
Currency Doubleday.
Sherif, C. W. 1976. Orientation in Social Psychology. New York: Harper and Row.
Vaughan, D. 1996. The Challenger Launch Decision. Chicago: University of Chicago Press.
Werhane, P. H. 1991. Engineers and Management: The Challenge of the Challenger Incident.
Journal of Business Ethics 10:605–616.
Werhane, P. H. 1999. Moral Imagination and Management Decision Making. Oxford: Oxford
University Press.
13 Viewing Trading Zones Developed to Advance Health as
Complex Adaptive Systems
Ann E. Mills, Mary V. Rorty, Lynn Isabella, and Donna T. Chen
The concept of a trading zone is particularly important today in the health-related

research world, where the seeming scarcity of publicly funded health research dollars
has resulted in calls for concrete deliverables: useful products and information that is
relevant to actual clinical or policy decision-making needs (Sung et al. 2003; Tunis,
Stryer, and Clancy 2003; Woolf 2008; Zerhouni 2005). The trading zone concept
highlights the importance of people coming together and working synergistically
toward a shared goal.
What happens when people work for a shared purpose can also be conceptualized
as a “complex adaptive system.” The idea of complex adaptive systems is especially
gaining attention in the organization and business literature, and many definitions of
complex adaptive systems have been offered. For instance, Plsek defines a complex
adaptive system as “a collection of individual agents that have the freedom to act in
ways that are not always predictable and whose actions are interconnected such that
one agent’s actions change the context for other agents” (Plsek 2001, 312–313). Uhl-
Bien and colleagues speak of complex adaptive systems as “neural-like networks of
interacting, interdependent agents who are bonded in a cooperative dynamic by
common goal, outlook, need, etc.” (Uhl-Bien, Martin, and McKelvery 2007, 299).
While the second definition emphasizes that a common goal is shared by those in
a complex adaptive system, both definitions highlight the interdependence of people
or institutions in such systems and the instability of the interactions of persons or
institutions involved. But this instability need not result in inappropriate outcomes.
The interactions of people in a complex adaptive system might instead result in
positive outcomes that may be surprising or innovative (Plsek 2001). But because
interactions are hard to predict in complex adaptive systems, outcomes too are hard
to predict.
The concept of complex adaptive systems explicitly recognizes that human beings,
even those engaged in cooperative enterprises, have individual motives and values.
266 Ann E. Mills, Mary V. Rorty, Lynn Isabella, and Donna T. Chen
Differences in motives and values will affect how the system functions and so can
affect the outcomes produced, just as the system itself will affect individual and insti-
tutional behavior. Thus, people working together in trading zones should be consis-
tently monitoring the goals, the values, and the outcomes of a trading zone. Goal,
values, and outcomes will affect interactions because they provide feedback that affects
the motivations and behavior of individuals involved in the trading zone. This, in
turn, will affect future interactions as well as outcomes. Paying attention to these
features of a trading zone can be particularly important to its success. This is particu-
larly true when a trading zone has been created to promote activities and to advance
goals that important stakeholders within it, and at times outside of it, consider to have
a fundamental normative dimension, as is the case for trading zones that come
together for the purpose of advancing human health and well-being through inter-
disciplinary research or action.
In what follows, we highlight certain features of trading zones by viewing them as
a type of complex adaptive system. Specifically, we focus on the goals and values of
trading zones and the individuals that constitute them, and show how these features
fundamentally affect the trading zone’s evolution, and thus its perceived or actual
success or failure. We argue that viewing trading zones as complex adaptive systems
highlights the need for consistent monitoring and explicit managing of a trading
zone’s goals and values to help it evolve toward its overarching goals and produce
appropriate outcomes, which is particularly important when fundamental normative
dimensions are at stake. We conclude by identifying some strategies to this end.
Trading Zones
Peter Galison uses the concept of a trading zone to explain how scientists and engi-
neers from different disciplinary cultures manage to collaborate across apparently
incommensurable paradigms. Collaborating across disciplines often involves the
development of a jointly understood pidgin or creole between individuals or groups
of individuals, which allows them to communicate meaningfully. Trading zones occur
when there is a need for interdisciplinary collaboration in the pursuit of a goal that
cannot be reached by collaborators within a single discipline. During World War II,
for example, the military, engineers, and scientists needed to collaborate across seem-
ingly incommensurable disciplinary paradigms in order to develop the weapons
needed to defeat the enemy (Galison 1996).
Trading zones can emerge when individuals, organizations, or governments per-
ceive an opportunity or a threat, and cooperate toward realizing the opportunity
Viewing Trading Zones Developed to Advance Health as Complex Adaptive Systems 267
or avoiding the threat. Michael Gorman, a psychologist, and James Groves, a mate-
rials scientist, used the metaphor of hiking to conceptualize the goal and travel
process of a small trading zone they formed (Gorman, Groves, and Catalano 2004).
Their problem was to figure out a nanotechnology project a graduate student could
carry out that would address a global problem, but still be doable within the scope
of a master’s thesis. They addressed their project by asking, “What sort of global
problem could be addressed by building a nanotechnology platform that could be
further built upon by others?” The global problem became a distant mountain
range; the path toward it, a trail; and the graduate student’s project a bridge that
would make it possible for others to take the trail—or to diverge and create another
trail from the bridge. In their vision of a trading zone, rivers and streams represent
the technical or research problems those in the trading zone would have to resolve
in order to proceed to their goal, represented by hills (immediate goals) or moun-
tains (superordinate goals). But bridges over the rivers and streams will not be built
unless the various participants in the larger project could communicate in a mean-
ingful way that is adapted to the goal its participants seek to achieve (Gorman,
Groves, and Catalano 2004). The trail that participants follow—or create, as needed—
evolves as bridges are built or as unforeseen obstacles or opportunities are
encountered.
The utility of a trading zone resonates in today’s world, where complex questions
and difficult problems beg for interdisciplinary attention and increasingly scarce
resources force people to look for innovative ways to increase productivity. As men-
tioned earlier, one of the goals of the health sciences is increased relevance and
utility of research. A major new road map for reaching this goal calls for increased
collaboration among multiple disciplines. A specific subset, called “translational
research,” is an approach that brings basic scientists together with clinicians, non-
clinician health care providers, policy makers, and others to develop research initia-
tives that reach back and forth between the bench, the bedside, and society (National
Institutes of Health 2008; Tunis, Stryer, and Clancy 2003; Woolf 2008; Clinical and
Translational Science Awards 2008). These initiatives extend throughout the processes
of developing new therapies, testing them, and bringing them to market as well as
figuring out how to get them to people in need who are not reached by the market.
Because it should take patients and their needs into account in all stages of these
processes, translational research in health care ideally represents a “patient-centered”
approach to research which is consistent with the Institute of Medicine’s call for a
patient-centered health care system (Institute of Medicine 2001). As part of the system
developed to improve human health and well-being, translational medicine has
normative aspirations that differentiate it from other types of collaborations or trans-

lational research in other areas.
In a variety of areas, contemporary research is hard to put within the bounds
of a specific discipline, or of one governing scientific or practice paradigm. Much
of it relies on getting together a number of talented people from a variety of dis-
ciplinary backgrounds—physicians, scientists, engineers, researchers, social scientists,
and humanists—and coordinating their efforts toward a goal of common interest,
with or without a single designated leader. The Manhattan Project and the associ-
ated Manhattan Engineer District project in the mid-twentieth century are examples
of trading zones (Boyer 1985). These projects are among the most discussed and
controversial because of their goals and their context. But the models they represent
continue to be implemented in numerous governmental and privately funded proj-
ects (National Institutes of Health 2008; Tunis, Stryer, and Clancy 2003; Woolf
2008; Clinical and Translational Science Awards 2008; Yamada 2008; Pfizer 2008).
Trading zones are being developed at the team and organizational level, called for
by recent competitive funding made available for translational health-related research
(Birmingham 2002), and at the national and international level in such projects as
the President’s Emergency Plan for AIDS Relief (PEPFAR), which aims to combat
the HIV/AIDS epidemic through funded program development and collaboration
with fifteen affected nations (PEPFAR 2008). Some of the Gates Foundation initia-
tives represent projects to improve global health through research and innovative
collaborative program implementation (Bill and Melinda Gates Foundation 2009).
Such projects involve individuals from disparate professional and disciplinary back-
grounds and represent public-private partnerships across disparate types of
organizations.
Complex Adaptive Systems
A “system” has been defined as “a complex of interacting components together

with the relationships among them that permit the identification of a boundary-
maintaining entity or process” (Laszlo and Krippner 1998, 51). Systems are structured
in ways that allow the achievement of their goals. They can be of various sizes—small
and self-contained with few interconnections, or large and complex, consisting of a
number of interrelated subsystems. For example, a light switch can be thought of as
a simple, mechanical system: it consistently produces the same outcome, and when
it does not produce the expected outcome, a signal is sent (the light doesn’t go on),
indicating that something is wrong with the system and that it must be fixed before
the outcome can be expected again. A computer can be thought of as a complicated

mechanical system because there are a large number of system components and a large
number of interactions among them. Systems are characterized as complex when the
system cannot be fully understood through its individual components; a cake can be
considered a complex system, for example, because it cannot be understood by simply
identifying its ingredients, or components (Uhl-Bien, Martin, and McKelvery 2007,
302). When we bring people together to accomplish a goal, the system they create by
working and thinking together should be thought of as a complex system, because
the system they create cannot be understood by looking solely at the system compo-
nents. Human systems have at least one more important feature—the systems they
create are adaptive, meaning that simply by interacting together, people create the
system and can change it through their interactions. Thus, human systems are emer-
gent and are constantly evolving (Plsek 2001; Uhl-Bien, Martin, and McKelvery 2007;
Stacey 2003).
So thinking of the interactions of an interdisciplinary team or other collaborative
initiative as mechanical or complicated or even complex may lead system designers
and members to conceptualize its function in simplistic terms. Research in a new
direction or a collaboration that involves multiple disciplines, backgrounds, or cul-
tures is always subject to obstacles and digressions derived from the subject matter or
overarching goal. In actuality, because an interdisciplinary team or other collaborative
effort is a complex adaptive system, obstacles may arise that relate not to the scientific
project, but to its interpersonal preconditions.
Why Goals, Values, and Beliefs Are Important in Complex Adaptive Systems
Individuals have unique driving motives and values that are encapsulated within the
mental model they hold for any activity of which they are a part. These mental models
are formed, and further reinforced, through a system or institutional social climate,
common culture, and functional infrastructure. As humans engage in activities, they
reflect on those activities and on the reactions others have to their activities, and they
learn from these reflections. Their learning serves either to reinforce or challenge their
operant mental models. Ralph Stacey, an organizational systems theorist, uses the term
“single-loop learning” to refer to the type of learning in which people monitor the
outcomes and take corrective action if needed, but do not change the fundamental
assumptions built into the organizational system (Stacey 2003, 111–112). When
operant mental models are not reflected on or challenged, the learning gets incorpo-
rated into a human system’s existing infrastructure, social climate, and common
culture. Stacey uses the example of a budget shortfall, in which there is often a process
of discovering why the actual outcome is different from the envisioned outcome
(Stacey 2003, 174). Once the reason for the shortfall is identified, a choice among
corrective actions can be taken, but in single-loop learning, significant alternative
actions might not be considered because they are not consistent with the dominant
mental model. Gorman and Werhane (chapter 12 of this volume) show how accoun-
tants at WorldCom used a single-loop process to deal with a major budget shortfall;
instead of questioning their mental model of the company’s likely future profits, they
relied on the assumption that they could apply a questionable accounting Band-Aid
to weather what they believed had to be just a short-term problem.
But human beings also are capable of “double-loop learning,” in which their
operant mental models, mindsets, or frames of reference are called into question
(Stacey 2003, 111–112). In the case of the budget shortfall at WorldCom, the assump-
tions that allowed individuals to envision a different outcome could have been ques-
tioned; they could have asked themselves whether their mental models or mindsets
actually represented reality (Stacey 2003, 175). For example, during a budget shortfall,
it may not be enough simply to cut costs in order to generate the desired revenue. It
may be that a budget shortfall represents a real change in patterns of consumer
demand and that a new product needs to be introduced. In this way, new goals might
be envisioned. Since human beings are capable of double-loop learning, their goals
and values are open to change, which allows for the possibility of creative, innovative,
emergent outcomes.
However, for the same reason, humans are not always predictable in their thoughts,
attitudes, and behaviors. Through their ability to accommodate unpredictable interac-
tions, complex adaptive systems are open to change. They can, and do, evolve. As
leading systems thinkers understand, one very small change can lead to a vastly dif-
ferent outcome over a relatively short period of time (Ray 1992). For instance, an
expert in a multidisciplinary collaboration may come to believe that the goal of the
trading zone cannot be accomplished in the envisioned time frame, or that she is not
able to contribute fully under existing conditions. Whether or not she communicates
this subtle change in belief to others, this change of mind will undoubtedly affect her
interactions with others and, for example, may ironically lead her to falter in timely
communications with others. On the face of it, this failure to communicate in a timely
fashion may represent a small change within the trading zone. But these interactions,
over time, might lead to a new reality in which the goals do become unattainable in
the specified time frame, or might lead to others being influenced by the expert’s belief
that the goal of the trading zone is unattainable. And so the trading zone may be
forced to reframe its goals, or in the extreme case it may fall apart. Because people’s
thoughts, attitudes, and beliefs are unpredictable and yet are vital to how human
organizations function, trading zones are best viewed as adaptive systems, rather than
as simple, complicated, or complex systems.
As a complex adaptive system evolves, the manner in which it evolves can be con-
sistent with and support its goals, or can diverge from and hamper evolution toward
its goals. In an example of the former, a complex adaptive system might find ways to
function more efficiently and effectively. It might evolve to create more innovative
outcomes. Complex adaptive systems designed to take advantage of the creative pos-
sibilities of double-loop learning might also experience evolution in their driving goals
and values. Consider the development of imatinib mesylate (also called Gleevec® or
STI571), a case frequently cited as demonstrating successful translational research
(Birmingham 2002).
Gleevec® was approved by the U.S. Food and Drug Administration (FDA) in 2001
for use in the treatment of some forms of adult and pediatric chronic myelogenous
leukemia (CML) (National Cancer Institute 2001). The development of Gleevec®
serves as an indicator for the future for other multidisciplinary teams involved in
researching other molecular-targeting cancer drugs because it shows that molecular
targeting can work if an appropriate target can be identified (National Cancer Institute
2002).
Three clinical trials demonstrated complete remission of CML in a large percentage
of the patients treated with Gleevec® and very limited side effects experienced by
patients (Druker et al. 2001). These outcomes were not expected for the early stage of
clinical research in which they were observed, yet were noticed as dosages were
increased. Because the trading zone developed for this effort included individuals who
could recognize this early signal and rapidly shift gears to take advantage of the oppor-
tunity presented, the driving goals of the trading zone changed as the implications
associated with increased dosages were understood. The initial goal of proving the
effectiveness and safety of Gleevec® were replaced with the goal of getting Gleevec®
approved for use by the U.S. Food and Drug Administration (FDA) as soon as possible
so that other patients suffering from the same disease might benefit. And in fact the
FDA gave their approval for the use of this drug after only a two-month review
(National Cancer Institute 2001).
But while changes might signal important positive evolution for the system, as was
the case with Gleevec®, it is also possible that these changes are undesirable. For
example, it is nearly always the case that when the safety and efficacy of a new com-
pound, therapy, or product are demonstrated, the developers want to get it approved
for use as soon as possible and keep it on the market for as long as possible under
exclusive patent. But there have been numerous cases when the inappropriate or
untimely embrace of the goals associated with commercialization have resulted in
falsified or misleading data being presented to, or withheld from, the FDA or society
(FDA 2008; Gilhooley 2007). In such cases, the health-related value initially governing
the formation of the project becomes less salient, and the participants lose sight of
their initial normative impetus.
It is often the case that goals and values are similar. If the goal of a system or insti-
tution is the delivery of high-quality health care, then the values associated with the
system should reflect this goal. But it is also important for trading zone participants
to understand that goals and values are not the same thing. Individuals, organizations,
and nations can have similar goals but differing values. Nations often share the goal
of eliminating hunger and disease, but differing values may make the success of a
trading zone aimed at attaining these goals problematic. As an example, in 2003 the
United States through PEPFAR initially committed $15 billion over five years to
combat the global HIV/AIDS epidemic in fifteen nations through global multi-
disciplinary, multiorganizational collaboration in program implementation. Goals for
combating the disease include both treatment and prevention (PEPFAR 2008). In
effect, PEPFAR represents a trading zone where different people of different cultures
and backgrounds are working together to solve a problem that threatens the survival
of many nations.
The nations that were chosen for inclusion in the initiative had similar goals to
those of the United States with regard to combating the HIV/AIDS epidemic. But
because many of the nations that wanted to participate in this initiative did not share
the values reflected in PEPFAR as conceptualized under the Bush administration, there
has been ongoing controversy. Brazil, which has a large HIV/AIDS population and has
worked hard to ensure that its people get the treatment they need, declined to further
participate in the initiative because PEPFAR requires that groups receiving funding
publicly oppose and condemn prostitution. Brazil feared that condemning commercial
sex workers, as well as other high-risk groups like injection drug users and men who
have sex with men, would hinder the country’s efforts to fight the disease by increas-
ing the stigma of the disease, encouraging concealment and discrimination (Kaiser
2005).
In addition to treating those with HIV/AIDS, one of the goals of the initiative is
preventing the spread of HIV/AIDS. This goal is also shared by the nations partici-
pating in PEPFAR. To this end, in 2003 Congress mandated that 20 percent of the
funds allocated should be used in ways that prevent the spread of HIV/AIDS. But
Congress also stipulated that at least 33 percent of that allocation had to be spent
on programs promoting abstinence until marriage. As the Institute of Medicine noted
in its evaluation of the progress of PEPFAR, abstinence-until-marriage programs
hamper prevention activities that are evidence-based, and are sometimes incompat-
ible with the epidemiologic and cultural context of some of the nations in which
they might be most needed (Institute of Medicine 2007, 113). In other words, even
though the stated goal of combating the spread of HIV/AIDS was shared among
nations, including the United States, the trading zone that emerged in the Bush
administration did not achieve its goal efficiently or effectively because of the nations’
conflicting values.
Furthermore, in order to capitalize on the potential benefits of single- and double-
loop learning, trading zones must allow participants to communicate and interact
freely with one another. This is how trading zone participants develop a creole, which
is necessary to “cross a bridge” in order to arrive at a goal. But the very interactions
needed to produce the creole might cause the system to evolve in ways that do not
support the system’s goal and its associated values, as in the above example of an
expert who no longer thinks a goal is attainable but does not communicate this belief
to others in the trading zone. Communication of this belief early on could have
allowed other trading zone participants to adjust to her new level of participation, or
to find someone else to fulfill the functions that they had originally desired and
expected of her. Lack of communication, as described earlier, could lead to the unex-
pected slowing of travel toward the goal or breakdown of the zone altogether. So on
the one hand, we want to constrain the system from evolving in undesirable ways,
but on the other hand, we want to allow it to develop an appropriate creole and to
evolve in desirable ways. The necessary ingredient is maintaining communication
about the goals and values of the trading zone and how these goals and values compare
to the driving motive, values, and mental models of the trading zone’s important
stakeholders—both those who are participants in the trading zone and those who are
outside of, but nonetheless important to, the trading zone. In any ongoing process
with changing conditions, positive evolution depends upon consensus and a commit-
ment to the values governing the shared enterprise from those interacting in the
zone. This consensus will provide a context, determine the social climate, and support
the development of an appropriate common culture for the interactions of those
in the zone, providing a constraint on otherwise inappropriate interactions. But
because the goals may change as the trading zone evolves, the consensus on values
and the identification of intermediate outcomes should be continually monitored and
explicitly managed.
Explicit Monitoring and Managing a Trading Zone’s Goals and Values
A consensus on, and commitment to, a trading zone’s goals and values will provide
a constraint of sorts on otherwise inappropriate interactions and evolution so long as
this consensus and commitment are continually monitored and managed. This is done
in part by aligning a trading zone’s infrastructure, social climate, and common culture
toward these common goals and values, assuming consensus and commitment are
achieved.
The activities associated with aligning a trading zone’s infrastructure, social climate,
and common culture toward agreed-on goals and values include, as we have said,
explicitly formulating and communicating a consensus on goals and values. But
the goals and values of a trading zone must be interpreted, continually examined,
and continually reinforced. This requires attending to the mental models held by
individual participants of the trading zone, as well as the mental models they
share, throughout the processes of single- and double-loop learning. Because a
zone’s infrastructure can enable the zone or hinder it, a third activity is necessary:
the goals and values of the zone must be implemented in the trading zone’s sup-
porting infrastructure. In this way, the social climate and common culture of a
trading zone will be created and reinforced. Moreover, because outcomes feed back
to influence the evolution of a system, outcomes will have to be monitored for
how well they align with the stated goals and values embedded in (and promul-
gated by) the trading zone. And as we have said, this is particularly important
when outcomes have a normative dimension, as in the case of translational research
in health care.
Trading zones are diverse. Some are large, some are small. Some will be supported
by a funding sponsor, some will not. Some will be formally created by an outside
person or entity, some will be spontaneous. So it is impossible to point to a single
individual or a person in a specific position as being responsible for managing and
monitoring the internal dynamics and outcomes of a trading zone. And when there
is no clear leader, these activities become, to some extent, the responsibility of every-
one who participates in the trading zone.
Sometimes formulating the goals and values of the zone and reaching a consensus
on them will require trading zone participants to negotiate among themselves. Other
times, the goals and values of the zone may be imposed on trading zone participants;
in these cases, if the imposed goals and values are incompatible with those of partici-
pants, the participants must either accept them or self-select out of the zone. Other-
wise, they may be asked to leave the zone.
Both trading zones and complex adaptive systems often have more than one goal
and more than one value. Because these goals and values can sometimes conflict, it
is particularly important to be clear about the primary goal of the trading zone, its
immediate goal, and the primary value of the system. Otherwise, there might be con-
fusion, frustration, or even anger about what a trading zone or complex adaptive
system is meant to achieve, when it is meant to achieve it, and how it is meant to
achieve it. As an example, many organizations fighting the HIV/AIDS crisis were
furious at the initial PEPFAR requirement that they denounce commercial sex work
in order to receive funding, because this requirement confused values with goals and
was inconsistent with the stated goal of PEPFAR. It called into question the actual goal
of PEPFAR: was it to promote a set of values, or to combat the HIV/AIDS epidemic?
Even when goals and values are shared among trading zone participants, they might
be interpreted differently. All participants might share the value of “excellence” but
be uncertain of what the term means in a given situation. Does it mean that all the
resources of a trading zone must be of a similar standard? Does it refer to excellence
in the methodology used to cross a paradigm bridge? Or avoiding errors or accidents
in the tasks associated with a particular subordinate project? As stated earlier, the
interpretation of the goals and values of a trading zone can either be negotiated among
its participants or imposed on them. Even if participants in a trading zone can arrive
at some sort of consensus on values, these values may be interpreted differently,
causing confusion and leading people to stop working together as a system, or allow-
ing the trading zone to evolve in undesirable ways. Moreover, this consensus must be
viewed as a flexible, rather than a fixed, interpretation, constantly reexamined or
revisited as the common project evolves, the context changes, or new participants
enter.
Here again, the concept of the trading zone is a useful one. As in a literal trading
zone, individuals with different backgrounds and a mixture of objectives meet in an
agreed-upon common area with a specific and often quite restricted purpose in mind,
with a common agreement to abide by the ground rules, a mutual understanding of
the boundaries of the activities to be carried out in that space, and a language, however
truncated—a trade pidgin or creole—whose implications are sufficiently transparent
to all participants.
Like any complex adaptive system, a trading zone needs some sort of infrastructure
that includes the resources of the system. The infrastructure will depend on the nature
of the collaboration; it may be as simple as a room with a computer, or it may
be large and complex, as when organizations or nations collaborate together, as in
the PEPFAR example. Regardless of the size and extent of the resulting infrastructure,
the infrastructure should support the goals and values of a trading zone rather then
inhibit them. The responsibility for managing these resources and developing the
infrastructure will depend on how large the trading zone is, where it is located, and
its fundamental nature. For instance, we can conceive of two experts interacting to
realize an opportunity who have no affiliation with other persons or entities. In this
case, managing the resources and developing an infrastructure that is appropriate to
the zone’s goal will be the responsibility of one or both of these people. But we can
also envision a large trading zone, such as PEPFAR, where there are designated roles
for managing the trading zone’s resources. In these sorts of cases, where an outside
person(s) or entity is able to affect the zone’s infrastructure, there is a potential
problem because the outside person(s) or entity must understand the nature of the
trading zone and its goals and values sufficiently in order to enable the trading zone
to achieve these goals and values.
So it is not enough that trading zone participants reach a consensus on the goals
and values of the trading zone. Nor is it enough that trading zone participants con-
tinually interpret these goals and values. It is also important that the goals and values
of the trading zone are reflected in the zone’s supporting infrastructure. This is what
we mean by saying that the goals and values of a trading zone must be implemented
in its supporting infrastructure. Monitoring and managing the internal dynamics and
outcomes of the zone, as well as developing and maintaining an enabling infrastruc-
ture, will create a social climate and common culture which together can provide a
constraint on the system, preventing it from evolving in undesirable ways.
Conclusion
Humans are separated as individuals and yet united as societies. In our long history
we have sought and found ways to work together as couples, as families, as tribes
and nations, and, to a small but hopeful extent, even in international and trans-
national organizations. Human interactions are difficult but desirable and, for some
purposes, necessary. Yet specialization brings with it socialization and professional-
ization within disciplines that impedes cooperation, even as the depth of knowledge
and expertise available to practitioners of the various specialties open new horizons
for scientific advancement and human progress. Despite their difficulties, the inter-
disciplinary scientific and research programs of contemporary societies represent
collaborations across individual differences and professional isolationism. In drawing
on concepts from the literature on complex adaptive systems and trading zones,
we hope to warn such collaborators of some obstacles facing them and to encour-
age some optimism about the future of such collaborations. Speaking of the evolu-
tion of individual sciences, physicist and historian Thomas Kuhn, in his canonical
work The Structure of Scientific Revolutions (Kuhn 1962), suggested that we might
understand the development of a new area of science as a series of diachronic shifts
of paradigm. A given area of inquiry can be initiated as an exploration of a given
problem in its pre-science stage, then becomes “normalized” as a specific theoretical
framework is adopted—and finally fades in importance or interest as the implica-
tions of its initial framework are exhausted. Revolutions in a given science occur
when a different paradigm is gradually adopted to replace the normalized theoreti-
cal framework as it loses its capacity to generate interesting results. The history of
the development of a science, then, from this standpoint looks less like the linear
progression of one strand of thought, and more like a composite of interwoven,
overlapping strands, with few elements persisting throughout the process. Although
some terms persist—matter, atoms, earth, or research team—they are understood
and used differently, have different denotations and connotations, at different his-
torical moments.
Today, multidisciplinary collaborations such as translational research have emerged
as the new paradigm for research appropriate for the expanding technologies and
increasing demands of contemporary society. We know that the trails of trading zones
will evolve as bridges are built or obstacles and opportunities are encountered. But
trading zones can be complex adaptive systems, and acknowledging that fact forces
us to examine how interactions might influence the evolution of a trading zone.
Interactions might cause a trading zone to move toward its goals in a creative and
unpredictable manner, but they might do the opposite, causing a trading zone to
evolve in undesirable ways. Thus, the infrastructure, social climate, and common
culture must serve to support the potential for innovative and creative desired out-
comes, while also acting as a kind of restraint on the system to keep it from evolving
inappropriately. Because goals, normative values, and outcomes influence the interac-
tions of people in a trading zone, we conclude that these goals and values must be
actively managed; they must be explicitly formulated, interpreted, and implemented
in the infrastructure of a trading zone.
References
Bill and Melinda Gates Foundation. 2009. Global Development. http://www.gatesfoundation

.org/global-development/Pages/overview.aspx (accessed March 2, 2009).
Birmingham, K. 2002. What Is Translational Research? Nature Medicine 8:647.

Boyer, P. S. 1985. By the Bomb’s Early Light: American Thought and Culture at the Dawn of the Atomic
Age. New York: Pantheon.
Clinical and Translational Science Awards. 2008. Translating Discoveries into Medical Practice.
http://www.ctsaweb.org (accessed April 4, 2008).
Druker, B. J., C. L. Sawyers, H. Kantarjian, D. J. Resta, S. F. Reese, J. M. Ford, R. Capdeville, and

M. Talpaz. 2001. Activity of a Specific Inhibitor of the BCR-ABL Tyrosine Kinase in the Blast Crisis
of Chronic Myeloid Leukemia and Acute Lymphoblastic Leukemia with the Philadelphia
Chromosome. New England Journal of Medicine 344:1038–1042.
FDA. 2008. FDA Takes Action against Compounded Menopause Hormone Therapy Drugs. FDA
News. Januray 9. http://www.fda.gov/bbs/topics/NEWS/2008/NEW01772.html (accessed May 10,
2008).
Gilhooley, M. 2007. Vioxx’s History and the Need for Better Procedures and Better Testing. Seton
Hall Law Review 37 (4):941–968.
Galison, P. 1996. Computer Simulations and the Trading Zone. In The Disunity of Science: Bound-
aries, Contexts and Power, ed. P. Galison and D. Stump, 118–157. Stanford: Stanford University
Press.
Gorman, M. E., J. F. Groves, and R. K. Catalano. 2004. Societal Dimensions of Nanotechnology.

IEEE Technology and Society Magazine 29 (4):55–64.
Institute of Medicine. 2001. Crossing the Quality Chasm: A New Health System for the Twenty-first
Century. Washington, DC: National Academies Press.
Institute of Medicine. 2007. PEPFAR Implementation: Progress and Promise. Washington, DC:
National Academies Press.
Kaiser. 2005. Politics and Policy: Brazil Refuses $40M in U.S. AIDS Grants to Protest Policy Requir-
ing Groups to Condemn Commercial Sex Work. http://www.kaisernetwork.org/daily_reports/
rep_index.cfm?hint=1&DR_ID=29729 (accessed May 12, 2008).
Kuhn, T. 1962. The Structure of Scientific Revolutions. Chicago: University of Chicago Press.
Laszlo, A., and S. Krippner. 1998. Systems Theories: Their Origins, Foundations and Develop-
ment. In Systems Theories and A Priori Aspects of Perception, ed. J. S. Jordon. Amsterdam:
Elsevier.
National Cancer Institute. 2001. FDA Approves Important New Leukemia Drug. http://www
.cancer.gov/newscenter/gleevecpressrelease (accessed May 10, 2008).
National Cancer Institute. 2002. Imatinib Mesylate (Gleevec™) Looks Even Better for Chronic
Myelogenous Leukemia. http://www.cancer.gov/clinicaltrials/results/gleevec-looks-better0302
(accessed May 14, 2008).
National Institutes of Health. 2008. NIH Roadmap for Medical Research. http://nihroadmap.nih
.gov (accessed April 4, 2008).
PEPFAR. 2008. What Is PEPFAR? http://www.avert.org/pepfar.htm (accessed May 14, 2008).
Plsek, P. 2001. Redesigning Health Care with Insights from the Science of Complex Adaptive
Systems. In Institute of Medicine 2001, 309–322.
Pfizer. 2008. Translation Medicine Research Opportunities. http://www.pfizer.com/b2b/

translational_medicine/translational_medicine.jsp (accessed May 14, 2008).
Ray, T. S. 1992. An Approach to the Synthesis of Life. In Artificial Life II: Studies in the Sciences of
Complexity, vol. II, ed. G. C. Langton, C. Taylor, J. Doyne Farmer, and S. Rasmussen. Reading,
MA: Addison-Wesley.
Stacey, R. D. 2003. Strategic Management and Organizational Dynamics: The Challenge of Complexity.
New York: Prentice Hall/Financial Times.
Sung, N. S., W. F. Crowley, Jr., M. Genel, P. Salber, L. Sandy, L. M. Sherwood, S. B. Johnson,

et al. 2003. Central Challenges Facing the National Clinical Research Enterprise. Journal of
the American Medical Association 289 (10):1278–1287.
Tunis, S. R., D. B. Stryer, and C. M. Clancy. 2003. Practical Clinical Trials: Increasing the Value
of Clinical Research for Decision Making in Clinical and Health Policy. Journal of the American
Medical Association 290 (12):1624–1632.
Uhl-Bien, M., R. Martin, and B. McKelvery. 2007. Complexity Leadership Theory: Shifting Leader-
ship from the Industrial Age to the Knowledge Era. Leadership Quarterly 18:298–318.
Woolf, S. H. 2008. The Meaning of Translational Research and Why It Matters. Journal of the
American Medical Association 299 (2):211–213.
Yamada, T. 2008. In Search of New Ideas for Global Health. New England Journal of Medicine
358:1324–1325.
Zerhouni, E. A. 2005. U.S. Biomedical Research: Basic, Translational, and Clinical Sciences. Journal
of the American Medical Association 294 (11):1352–1358.
14 Creating Trading Zones across Continents and Economies:
The Female Health Company
Mary Ann Leeper, Elizabeth Powell, and Patricia H. Werhane
Chapter 12 in this volume, by Michael E. Gorman and Patricia H. Werhane, borrowing

from Peter Galison (1997), defines a trading zone as a locus of communication, often
involving the development of a jointly understood “pidgin” or “creole” between
individuals or groups of individuals whose background or theoretical points of view
were vastly different and sometimes incommensurable. In this chapter we shall see
how trading zones are useful in commerce to produce some extraordinary results that
would otherwise not be possible.
The Female Health Company
To understand the evolution of the Female Health Company’s work in developing

trading zone strategies to enable the core operation of its business, we must first get
a picture of the company’s struggle to play a role in the fight against HIV/AIDS. The
story of the Female Health Company (FHC) starts in 1986, when three executives from
a large pharmaceutical company decided to step out on their own to build a health
care company. They set out to identify small health care companies that were strug-
gling due to lack of resources and management talent but that had existing products,
or products in development, with the potential to be the foundation for a strong
future.
To start, the three former G. D. Searle pharmaceutical executives, O. B. Parrish,
William Gargiulo, and Mary Ann Leeper, bought management control of the Wiscon-
sin Pharmacal company, located in Jackson, Wisconsin. They were intrigued with
the base business and the prospect of licensing in the yet-to-be-developed “female
condom” identified in Wisconsin Pharmacal’s product pipeline. At the time, AIDS was
newly recognized as a deadly disease in the United States, particularly among men
and women living in high-risk situations. And while male condoms had become a
household product sold in almost every drugstore in America for prevention of
282 Mary Ann Leeper, Elizabeth Powell, and Patricia H. Werhane
sexually transmitted diseases, nothing was available for women to use for this purpose.
It seemed an obvious next step to introduce a female condom. Women needed a
protection method that they could initiate, something that would be under their
control. Parrish, Gargiulo, and Leeper assumed that, like the male condom, the female
condom would soon be on the shelves of every store and widely purchased by sexually
active women.
The three senior executives now managing Wisconsin Pharmacal began to negoti-
ate with the female condom’s inventor, Danish physician Lasse Hessel, and to seek
the funding required to develop both the Wisconsin base business and what they
hoped would be their “gold coin,” the female condom. While building a management
board and taking the Wisconsin company public, Mary Ann Leeper, the COO of the
company, undertook the long process of developing the female condom to meet the
Food and Drug Administration (FDA) requirements and to gain final approval to
launch the product in the United States.
The Product and Its Launch
The female condom is made of a sheer, nonlatex, polyurethane material and is shaped
as a sheath with flexible rings on each end. Because the female condom transfers heat,
the plastic used is strong and impermeable to temperature changes, humidity, and
oil- or water-based lubricants. More importantly, the sheath cannot be penetrated by
various viruses and bacteria that cause sexually transmitted diseases (STDs) like tricho-
moniasis, the herpes virus, the hepatitis B virus, and the lethal HIV. The condom’s
uniqueness lies in the control it gives women to protect themselves against sexually
transmitted infections as well as unintended pregnancies.
The job for COO Leeper was not only the condom’s technical development but
also the marketing strategy for its introduction. The product handed over from the
inventor was basically a crude design. Leeper had to finish the product design, finalize
the materials to be used, and ensure that the right preclinical, clinical, and safety
studies were undertaken to a standard that would meet the U.S. FDA requirements.
She and her team assumed that the marketing strategy should be similar to that of
other consumer products that were sold in drugstores.
However, the development of the female condom and ultimate approval by the
FDA took over six years and at a cost that was magnitudes higher than originally
anticipated. And development was only the beginning of the process; the product
launch and marketing strategy were even higher hurdles to get past. Market research
results were mixed. According to Leeper, “For every complaint we got, there was an
Creating Trading Zones across Continents and Economies 283
equal amount of positive feedback. No one really knew how the consumer was going
to judge the product.”1 The product had many perceived benefits when compared
with other prophylactics and contraceptives. The device did not require health care
services to be fit, prescribed, refilled, inserted, or removed; it had no systemic (hor-
monal) side effects for the user; it did not require partner consent because it could be
inserted well in advance of sexual activity and therefore was less disruptive to inter-
course; it was not necessarily obvious to the male partner, and was odor and taste free.
There were, however, some obstacles noted: the female condom was aesthetically
unusual; some women had difficulty inserting it, while others complained of discom-
fort or reduced sensation; occasionally the condom became dislodged; and some
partners objected to its use.
Nevertheless, Wisconsin Pharmacal, later renamed the Female Health Company
(FHC), prepared to distribute the female condom in both the public sector and on the
commercial market. By the end of 1993, the female condom was available in the public
sector, where physicians and public health care providers were the customers who
would be responsible for recommending the product to end users. Broad-scale com-
mercial distribution was initiated nine months later.
To introduce the female condom to young women in the United States, the
company used the standard approach to new product introduction to consumers.
Advertising campaigns were based on the results of marketing research, with sup-
porting promotional efforts aimed at physicians, nurses, and health care providers.
Product branding and marketing targeted the idea of women’s empowerment and
positioned the product as an alternative to the male condom. The female condom
was sold in supermarkets and drugstores, shelved with other personal care
products.
The product launch to the public sector used a different approach, emphasizing
the education of public health workers and end users about the benefits of the female
condom. It progressed slowly but on track. Using a female condom was a new method
of STD prevention, so its introduction in a clinical setting had to be carefully designed
and implemented; an attractive advertising campaign was not enough. While the
public-sector program seemed to be working, the progress and results of the commer-
cial launch were a different story. Even the market research did not foretell what would
happen. Despite the hefty investment, the whole campaign failed. Within six weeks
of the national launch, Leeper realized that a struggle lay ahead and a new strategy
had to be developed. Young sexually active middle-class American women did not
recognize they were at risk of getting an STD, let alone the possibility of being infected
with HIV and dying.
The Public and Commercial Sectors
Based on more marketing research work in both the public and commercial markets,
the company shifted focus from mass advertising in the commercial market to
directly addressing the public health communities in several major metropolitan
areas where HIV/AIDS was recognized as a serious issue. The shift was away from
direct marketing to young women and toward bringing a prevention tool to those
who recognized they needed it, mostly to public health clinics that treated at-risk
women who were not protected by their male partners. The marketing strategy
changed from a focus on middle-class college-aged women to women with the
highest risk of sexually transmitted diseases. Along with the focus on the public
sector came a clear recognition that a whole new education process had to accom-
pany the introduction of the female condom. This new marketing strategy entailed
a shift of mindset. When the company originally developed the female condom for
over-the-counter sale to middle-class women, it was not clear that the primary cus-
tomers would be women who felt directly threatened by STDs and AIDS. This new
customer base forced the company to rethink their efforts as not merely producing
a product but developing a program. Thus, the strategy became a program that
combined product information and education outreach, working with all of the
stakeholders that would influence the female condom’s purchase and use. This real-
ization about the product’s customer base laid the groundwork not only for its
public-sector distribution in the United States but eventually for the introduction
of the female condom to the global public sector.
The plan introduced a whole new communication program—a step-by-step program.
The public sector campaign amounted to a “train the trainers” program designed to
teach the counselors so that they, in turn, could help their clients. This approach
generated acceptance, sales, and satisfied users. Still, the costs of development and
FDA approval had created large corporate debts. And FHC struggled to survive.
A New Way of Thinking
In early 1996, FHC (now headquartered in Chicago) was contacted by Daisy Nyamu-
kapa, manager of the HIV/AIDS Coordination Programme for Zimbabwe’s Ministry of
Health and Child Welfare. Over thirty thousand Zimbabwean women had signed a
petition demanding that the government bring the female condom into their country.
This was a gigantic market opportunity. The task for FHC became how to take what
they had learned by trial and error in the United States to work in a country halfway
around the world with a different culture and no resources. The plan was to move the
U.S. public-sector approach into the global market.
Here began a communications strategy beyond FHC’s imagination. Zimbabwe was
just the first of what ultimately became at least twenty-five countries identified as
having a huge need for this product and thus places to get the message out. Education
outreach programs had to reach the women who needed to learn about the female
condom and what it could do to help them protect themselves. FHC also had to find
a way to train counselors, fund the product, distribute the product within country-
specific systems, and reach out to those who needed the female condom the most.
The question became: How does a small Chicago-based company with very limited
resources mobilize an effort to reach the women at risk in the villages and rural areas
of sub-Saharan Africa?
The company’s challenges were multifaceted. Leeper, as COO, bore the responsibil-
ity for meeting these challenges. She needed to win funding for those countries most
in need of her product from international foundations, many of which questioned
the funding of products made by a for-profit publicly traded American company.
Leeper needed a set of delivery systems customized for different cultures and that
insured proper use of the product; and she had to worry about institutional barriers
to entry into countries that had reputations, reinforced by Transparency Interna-
tional’s annual corruption ranking, for rampant bribery and corruption. She also was
intent on protecting the intellectual property and patents FHC had garnered for the
female condom. To master these challenges, Leeper had to bring into question various
operative mental models, including her own training as an MBA and manager of a
for-profit Western company, and develop a strong moral imagination to think differ-
ently in a new context. Most importantly, Leeper had to develop a trading zone to
communicate across these perspectives.
Trading Zones and the Female Condom
To understand the strategy of trading zones as applied to female condom distribution

in sub-Saharan Africa, one must first acknowledge that the interrelationship of poverty,
lack of schooling, lack of food supply, nonexistent or very limited health care infra-
structures, and institutional barriers such as rampant corruption and very limited rule
of law created fertile ground for the spread of HIV/AIDS. In order to attack HIV, it
would take an integrated approach, not merely the initiative of a single company, to
even begin to have an effect. The Female Health Company came to this realization in
part because of its experience in the United States, and began to reach out to various
stakeholders and groups in-country who would play a role in the education and
product integration programs needed for a successful introduction. International non-
government agencies (NGOs), country NGOs, global health agencies such as the World
Health Organization, and other United Nations agencies rallied to the effort.
Part of the implicit foundation for developing this trading zone began with a
prioritization of a set of agreed-upon values—in this case, the values of human life,
health, and disease control. These were values shared across governments, NGOs,
social networking organizations, aid and funding organizations, pharmaceuticals and
biotech companies, as well as the patient populations to be served. While it is often
the perception that for-profit ventures are only in the business of making as much
money as possible, the mission of almost every pharmaceutical company belies that
priority. Part of any for-profit health care venture is, by definition, the goal of health,
disease control, or palliative care. Otherwise, these businesses are not in the business
of health care. Although these values were often merely implicit, they served as the
basis for communication across theoretical and cultural barriers. The FHC strategy
was to use these values as part of the commitment from various stakeholders and
to draw upon the urgency of the pandemic as the starting point for developing a
shared superordinate goal that would motivate communication and cooperation
across various disciplinary, cultural, and economic boundaries to set up goals for
this venture.
Next came the search for funding for country-specific programs. Major national
and international donor agencies, such as the United States Agency for International
Development (USAID), the World Bank, the European Union, and the Netherlands
Bank, as well as private donors such as the Gates Foundation, were invited to come
to the table to supply or support the resources required to implement country-specific
HIV prevention programs which included the female condom. FHC spent many years
sorting out the details of the public/private partnerships and forming the alliances
required to implement a full-scale education outreach and distribution program within
country to reach the women on the streets of their villages. Each country had its own
culture, its own set of stakeholders, and its own priorities, as did each of the alliance
groups. FHC’s female condom operations had to take all of these aspects of the NGOs,
donors, country governments, and the company itself into consideration when devel-
oping country-specific strategies.
Rather than calling this venture a “trading zone,” Leeper called it an alliance and
a joint venture, which depended on close interrelationships between stakeholder
networks. But the venture functioned as a virtual trading zone, because each set of
actors (NGOs, donors, governments, and FHC) had to communicate across different
languages and priorities to focus on the goal of reducing the numbers of HIV/AIDS-
infected women. Secondly, she renamed what she was doing a “program” in order to
reinforce the interactions between disparate parties. By engaging in a program rather
than merely marketing, funding, or using a product, FHC shook off the aura of being
a foreign multinational that was merely marketing a product with the sole goal of
large sales, and thus large profits.
Then, by working “on the ground,” village by village, with local NGOs and other
social workers, the program (and thus FHC) has been able to sidestep government
interference; in fact, working with the governments was critical to the program’s
success. Because FHC’s program is a socially valuable one focusing on control of a
pandemic, the company is never approached for kickbacks in order to operate locally
or nationally in the approximately 100 countries where the program operates.
The biggest challenges for FHC continue to be funding; communication across
various cultural traditions, some of which preclude women from making their own
choices; and traditional practices of indigenous medicine based on the belief that HIV
can be prevented or cured by herbal medicines, incantations, and other such practices.
All of these are global challenges to health delivery organizations and are not unique
to sub-Saharan Africa or rural communities in Asia.
Despite these and other difficulties, FHC now operates in over 100 countries,
working in various alliances with their unique product to reduce infection rates
among sexually active women and their yet-to-be-born children. This continuing
success is attributed to the formulation of a stakeholder alliance model, redrawing
traditional stakeholder maps and aligning a wide network of values-based responsi-
bilities for the program. In other words, the company has created a highly successful
trading zone.
Today, the overall approach is working because of the coming together of very
different groups with an agreed values orientation to fight a very specific cause: the
HIV pandemic. Today, thousands of women at risk who live in rural communities in
economically underdeveloped countries have access to, and use, the female condom
as their method of protection against HIV.
The failure of the female condom launch in the U.S. market shows that the intro-
duction of a product that addresses a critical social need can only be truly successful
when all stakeholders come together with an integrated and agreed-upon approach.
FHC had assumed that its original target groups of middle-class women and female
college students would recognize the value of the female condom when it was intro-
duced in the United States. But, because these groups didn’t feel they were at risk of
HIV or STDs, they did not recognize the product’s value. FHC then had to seek out
those who needed the product and determine how best to align these needs with
funding and communication—making the female condom available to those who
needed and wanted it.
Note
1. Quotations from Mary Ann Leeper are from Elizabeth A. Powell and Gerry Yemen, “The Female
Health Company (A)” and “The Female Health Company (B),” case studies written for the Darden
School of Business, University of Virginia, UVA-BC-0146 and UVA-BC-0183.
Reference
Chicago Press.
15 Conclusion: Future Research on Trading Zones and
Interactional Expertise
Michael E. Gorman
This volume is not an end but a beginning. The 2006 workshop on Trading Zones,
Interactional Expertise and Interdisciplinary Collaboration was followed by two work-
shops on Studies of Experience and Expertise (SEE), held by Harry Collins and Rob
Evans at Cardiff. These workshops encouraged the development of a community of
scholars who would continue work in SEE.
In addition, the psychology of science, long considered a minor area in science and
technology studies (STS) and in psychology, is now achieving status as an interdisci-
plinary field, with its own society and journal, the Journal of Psychology of Science and
Technology. The study of scientific and technological expertise is one of its topics of
interest (Gorman 2008).
This section on future directions is adapted from a presentation Harry Collins gave
on my behalf at the latest (2008) SEE meeting. There is no trading zone between SEE
and the psychology of science at present, no creole or even shared jargon. There are
advantages to having different perspectives on, and approaches to, the study of exper-
tise. A trading zone creates a space for exchange without inhibiting this diversity of
methods and perspectives. A new, united field might eventually emerge, or separate
research communities might continue their own approaches. The point is to com-
municate while preserving diversity of knowledge, perspectives, and methods. A
trading zone is essential because research on the psychology of expertise is relevant
to SEE, and vice versa.
As a first step toward creating a trading zone, let us consider methodological
approaches from the psychology of science that might be useful for studying trading
zones and interactional expertise; that is, fitting the research in this volume into the
categories, and if the categories do not fit, transforming the categories. These catego-
ries were originally proposed by Kevin Dunbar and J. A. Fugelsang, who based them
on an analogy to biological research (Dunbar and Fugelsang 2005).
290 Michael E. Gorman
In Vitro
In vitro studies correspond to laboratory experiments in psychology conducted via

artificial tasks with participants who are not domain experts. One could, for example,
turn a game like Civilization IV into a task where the survival of all civilizations
would require the formation of trading zones. In Civ IV, each player manages a
civilization over an extended interval of time, and has the choice to either cooperate
or compete with other players managing different civilizations. Although the game
does not simulate a tightly coupled system where a collapse can only be avoided
by cooperation, such coupling could be achieved through an unequal distribution
of vital resources over the simulated globe and through phenomena like global
warming. These coupling variables could be manipulated to see their effects on
trading zones.
Interactional expertise could be incorporated by having additional participants
with no prior knowledge of the game learn about it by conversing with those playing.1
Under what circumstances would these outsiders be able to acquire enough knowledge
to pass as a gamer in serious conversation about strategies, preferred styles of play,
and the like? Would they be able to pass as a gamer in a Turing test?
An alternative is to do a simulation in which participants have to occupy different
roles. Consider Nanosim, a simulation of the National Nanotechnology initiative that
I have used in several of my classes at the University of Virginia, including an NSF-
supported course on Societal Dimensions of Nanotechnology.2 Students are placed in
groups corresponding to the following roles:
• Congress,
• Funding agencies like DARPA, NSF, and NIH,
• Companies like IBM and entrepreneurial start-ups,
• University laboratories,
• NGOs like the Project on Emerging Technologies and ETC,
• A newspaper that reports to all the other groups.
Students representing laboratories and companies make choices about what nanotech-
nologies to create, and collaborate or compete as they move up a technology tree to
acquire the rights to increasingly complex technologies. Those representing Congress
supply funding, depending on how the research is justified, and can create rules to
ensure fairness, societal goals, etc. Students representing NGOs can use a variety of
strategies to encourage or block technologies. In order to create new technologies,
Conclusion: Future Research on Trading Zones and Interactional Expertise 291
participants have to trade resources, intellectual property, and time. Outside events
can also be introduced to alter the simulation.
In order agree on goals for the NNI and make progress toward grand challenges like
a cure for cancer or a space elevator, students have to form trading zones across their
different roles and teams. The bar is low for students: they spend only a short time
in their roles, they are not ideologically wedded to them, and the stakes are small—
they will not lose jobs or status if they fail. Still, Nanosim incites passionate arguments
around who should be included and left out of trading zones; some trading zones fail,
and some turn into alliances that allow combinations of groups to discover new tech-
nologies that no group alone could achieve.
Nanosim is currently an educational simulation, but it could be turned into an
experimental space by taking multiple measures relevant to trading zones during the
course of the simulation, and exploring variables that might enhance group members’
adherence to their roles.
The advantage of in vitro studies is that they allow control and manipulation of
variables. The disadvantage is that they have what the psychologists call low ecological
validity—they involve tasks that are only analogous to real problems, and participants
are not separated by deep expertise and cultural divides.
Ex Vivo
Ex vivo studies, on the other hand, involve actual scientists or engineers, working on
artificial tasks. An example is the imitation game described in chapter 4 by Robert
Evans and Harry Collins. Along with an actual gravitational wave physicist, Collins
was asked the sorts of questions an expert in this domain ought to answer, and experts
were unable to identify Collins as the one who was not a real physicist. Further studies
of this sort could be done in other domains of science to see how well the findings
generalize.
Sub Specie Historiae
Fine-grained historical case studies can also be analyzed to see if interactional expertise
was acquired by any participants and whether and how trading zones were formed.
But the right kind of fine-grained details have to be recorded. J. Robert Oppenheimer
moved easily across scientific boundaries as he encouraged work on the atomic bomb
(Thorpe and Shapin 2000). He could not have done original research in all the areas
of expertise involved, though his knowledge of physics was helpful across the board—
so he might have acquired interactional expertise while working at Los Alamos. While
scientists and engineers often keep notebooks of their discoveries and inventions, the
only expertise acquisition diary I am aware of is that of Jeff Shrager (Shrager 2005).
Peter Galison has provided substantial historical evidence of the development of
trading zones. Yet, as far as I know, there is no historical study of interactional
expertise.
In Vivo
In vivo studies require observation of experts “in the wild,” working on the sorts of
problems they would actually confront. In this volume, for example, Lekelia D. Jenkins
(chapter 8) interviewed and observed marine scientists, fishers, and other stakeholders
in the TED trading zone, and showed how changes in the zone can be graphed. Erik
Fisher (chapter 10) worked as a participant observer in a nanotechnology laboratory.
And Jeff Shrager (chapter 6) took participant observation to a new level by becoming
a molecular biologist and designing tools to encourage the formation of trading zones.
This kind of participant observation raises the concern that the researcher will “go
native” and lose the ability to analyze the situation from an outsider’s perspective. All
of the authors in this volume were able to keep their anthropological distance by using
rigorous analytic methods, including reflective diaries and branching tree diagrams.
They also discussed their work frequently with members of scholarly communities in
STS and cognitive science. In addition to observation, in vivo studies can involve the
following:
• Interviews: Observation and participation can be supplemented by interviews with
participants in a trading zone. It is best if these interviews are done while the trading
zone is active. The major weakness of retrospective interviews is that human memory
is constructive, so informants may reconstruct their actions as they think they ought
to have happened (Ericsson and Simon 1984).
One interview strategy that should be particularly useful for the study trading zones
is the critical incident technique (Zsambok and Klein 1997), in which experts are asked
to recall in detail a critical situation where they had to apply their expertise, and as
much as possible about how they applied it and what resulted. Those involved in
collaborations could be asked to describe critical incidents in the development of the
trading zone, including successes and failures. Researchers could prompt for evidence
of an emerging creole and whether any members of the collaboration possessed
interactional expertise. (The prompts would have to be done carefully, so as not to
bias the interviewee.) Those who were purported to have interactional expertise could
then be tested via the imitation game, and also interviewed.
• Surveys: Questions could be derived from observations and interviews and used to
assess how well results from fine-grained in vivo research generalize to other science
and engineering collaborations.
• Documents: Grant proposals, drafts of articles, e-mails, and other documents should
also be included in the analysis, where possible. Here the in vivo data converges with
the sort of data used in sub specie historiae analyses.
In Silico
Dunbar and Fugelsang (2005) use the term “in silico” to refer to computational
models. It would be possible to model how trading zones shift from being coercive to
cooperative, from homogeneous to heterogeneous—or the reverse. These models are
akin to in vitro studies in that they have a high level of control and precision; one
can introduce and manipulate variables and see their effect in virtual space. Models
can potentially have higher ecological validity than experiments if they are linked
closely to solid in vivo data—like the program David Gooding and Tom Addis (1999)
created to explore how Faraday could have taken different discovery paths, at a fine-
grained level of detail, experiment by experiment, and supplemented by replications
of some of the actual experiments. The great weakness of computational models is
that they are often confused with reality, and they always embody tacit assumptions
of the modeler. Therefore, models have to be developed in close contact with data,
and their assumptions stated clearly.
Turning Framework into Theory
This empirical work would contribute to turning the current framework into a theory
that made predictions. A good first step would be refining the current taxonomy of
trading zones. As Galison emphasizes in chapter 3, not every multidisciplinary interac-
tion is a trading zone. There must be a reason for distinct subcultures within science
and technology to trade by exchanging ideas, resources, and/or credit. These zones
could be initially classified using coercive/subversive and homogeneous/heteroge-
neous dimensions, which are not binary, but matters of degree.
Jenkins demonstrates in her chapter how empirical work on trading zones can lead
to refinement of the theory; her study includes the emergence of a second trading
zone to recruit allies for the solution that emerged from the first trading zone, and
considers the addition of boundary organizations to the framework. She ends up

depicting the trading zone taxonomy in a way that places more emphasis on trajec-
tories and their dependence on shared language.
The research methods outlined above could be used to explore these trajectories.
Empirical case studies—including historical ones—could be graphed and compared;
from those graphs, models could be developed to describe different trading zone tra-
jectories, and predictions derived from the models. These predictions could, in turn,
be tested both in vitro and in vivo.
The trading zones concept also has limits, and these limits need to be defined more
rigorously. Could most types of interdisciplinary exchange in science and technology
be described as trading zones, on a kind of continuum where some of these exchanges
fit the concept perfectly and others barely at all? Could the same be said of trading
zone subtypes? Or should the concept have definite limits—this kind of exchange is
a trading zone, this other kind is something else entirely?
The trading zone mediated by a boundary object is a particularly promising area
for future research because the term “boundary object” is itself ambiguous. Sketches,
maps, and other kinds of visualizations can serve as “wordless creoles” (to adopt
Galison’s phrase); artifacts and systems should be able to as well. The Everglades, for
example, serves as a boundary object for those stakeholders trading around its future,
and each sees the Everglades in different ways—but this seeing is constrained by the
current state of the sociotechnical-natural system (see chapter 7). The use of dams can
be negotiated in the future, but the facts of their current numbers, placement, and
operation could act as a kind of visual creole—that is, all parties can go and observe
their operation, even if they disagree strongly about their utility and have different
levels of understanding of the technology. Questions like this can be resolved empiri-
cally, by looking at the role of artifacts and systems in achieving partial coordination
across trading zones.
Interactional expertise as currently operationalized is an all-or-nothing affair:
either one can pass an imitation game test, or one cannot. More detailed research in
how interactional expertise is acquired might also reveal meaningful stages in its
acquisition.
Another question about trading zones has to do with breadth. For example, I often
have to gain enough interactional expertise to help an engineering colleague supervise
a graduate student doing research into a sociotechnical system, but my knowledge is
limited to the task at hand, e.g., silver nanoparticles in water as an antimicrobial agent
(Gorman et al. 2008). Is interactional expertise limited to narrow domains, or can one
talk meaningfully about interactional expertise, say, in nanotechnology? One way of

answering this question would be to see if one could construct a meaningful Turing
test, identifying domain experts.
There may be a taxonomy of other types of partial expertise in the work of a com-
munity. It is possible for an interactional expert to gain tacit knowledge of the lan-
guage of an expert community. It may also be possible for another kind of expert to
gain tacit, hands-on knowledge of the research practices in a community without
being able to use the language. I have learned how to operate a telescope that is more
than a century old, giving me an experiential sense of what it was like to be an
astronomer in 1885 possessing a top-of-the-line instrument—but I cannot pass as an
astronomer in discourse with experts, or even with amateurs (Gorman 2006). This
kind of expertise is possessed by technicians who do hands-on laboratory work.
Application Areas
These same methods can be used to study the application areas highlighted in this
volume: service science (Gorman), business consulting (von Oetinger), environmental
management (Allenby), education (Mehalik), and ethics (Werhane, Leeper, and
Gorman). Obviously, the best approach is to use multiple methods that comple-
ment one another. For example, in the service science area, one could do in vivo
studies of several provider-customer relationships over time, combining observations
with critical incident interviews to determine how service professionals and clients
solve critical problems together; and one could conduct ex vivo experiments to
determine if the service scientists had become interactional experts in the client’s
domain.
The ethical perspective reminds us that one goal of the empirical research will be
to reach normative conclusions about best collaborative practices, and that best prac-
tices are always ethical practices. In particular, moral imagination provides a potential
solution to the kinds of values conflicts that create incommensurable barriers to col-
laboration (see chapter 12 by Gorman and Werhane). Here empirical research on the
effectiveness of moral imagination is essential, using a combination of the methods
outlined above.
The evolution of our sociotechnical system has taken us to the point where we are
exploring other worlds, creating astounding advances in health care—and at the same
time killing each other. As Walt Kelly wrote, “We have met the enemy and he is us.”
We have also met the opportunity, and it is us. The key is collaboration.
Notes
1. I find I can understand the strategies and goals in complex computer games my lads and their
friends play, and even talk about them with my lads at a fairly sophisticated level, without my
being able to play the games myself.
2. NSF grant EEC-0836648 from the Nanotechnology Undergraduate Engineering program.
References
Dunbar, K. N., and J. A. Fugelsang. 2005. Causal Thinking in Science: How Science and Students
Interpret the Unexpected. In Scientific and Technological Thinking, ed. M. E. Gorman, R. D. Tweney,
D. C. Gooding, and A. Kincannon, 57–80. Mahwah, NJ: Lawrence Erlbaum Associates.
Ericsson, K. A., and H. A. Simon. 1984. Protocol Analysis: Verbal Reports as Data. Cambridge, MA:
MIT Press.
Gooding, D. C., and T. Addis. 1999. Learning as Collective Belief-Revision: Simulating Reasoning
about Disparate Phenomena. Proceedings of the AISB Symposium on Scientific Creativity, 19–28.
Sussex: Society for the Study of Artificial Intelligence and the Simulation of Behaviour.
Gorman, M. E. 2008. Scientific and Technological Expertise. Journal of Psychology of Science and
Technology 1 (1).
Gorman, M. E., A. Wardak, E. Fauss, and N. Swami. 2008. A Framework for Using Nanotechnol-
ogy to Improve Water Quality. In Nanotechnology Applications for Clean Water, ed. N. Savage, M.
Diallo, J. Duncan, A. Street, and R. Sustich. Norwich, NY: William Andrew, Applied Science
Publishers.
Gorman, M. E. 2006. Thinking Like an Astronomer. In Organizations and Strategies in Astronomy,

ed. A. Heck, 419–428. Dordrecht: Springer.
Shrager, J. 2005. Diary of an Insane Cell Mechanic. In Scientific and Technological Thinking, ed.
M. E. Gorman, R. D. Tweney, D. C. Gooding, and A. Kincannon, 119–136. Mahwah, NJ: Lawrence
Erlbaum Associates.
Thorpe, C., and S. Shapin. 2000. Who Was J. Robert Oppenheimer? Social Studies of Science 30
(4):545–590.
Zsambok, C., and G. Klein. 1997. Naturalistic Decision Making. Mahwah, NJ: Lawrence Erlbaum
Associates.
Index
actor network theory, 182–185, 190 Bertelsmann, 237

Addis, Tom, 293 Bijker, Wiebe, 86, 184–185
Afferent, 112–115 BioBike, 72, 89, 117–122
Agency for International Development, 286 biocomputing. See computing
AIDS. See HIV/AIDS BioLisp, 89
Alexander, Annie Montague, 2 BioLite, 89
Allenby, Brad, 72, 99n31 biology, and statistics, 108–109
Alvarez, Luis, 30 biotechnology, 1, 136
American exceptionalism, 133–134 Bjorken, James, 32–33
American Telephone and Telegraph Company Bloch, Marc, 25
(AT&T), 258 Boisjoly, Roger, 247, 251
Anderson, John, 111 Boston Consulting Group, 2
Annales School of historiography, 25 boundary objects, 2–3, 8, 12, 16, 46, 82–84,
anthropogenic earth 86, 121, 165, 175, 189–191, 294
accelerating change, 125, 127, 132, 137, Braudel, Fernand, 25
139–145 Braverman, Harry, 25
complexity, 125, 127, 130, 136, 139–146 business world
and personal authenticity, 145–148 corporate culture, 231–232
radical contingency, 125, 127, 131, 137–145 likelihood of trading zones, 239
Apple Computer, 237 non–trading zone exchange, 236, 239
Aral Sea, 127–128, 131–132, 148, 151 scientific interdisciplinarity in, 233–234
ARPANET, 16, 87–88, 191 structures for exchange, 238
Atomic Energy Commission, 48 bycatch, marine, 72, 78, 157, 160–161.
See also turtle excluder device (TED)
Bagian, James, 262n2
Bainbridge, W. S., 138 Callon, Michael, 183–185
Baird, Davis, 1 Campbell, Joseph, 99n27
Ball Aerospace, 210 Carlson, W. Bernard, 185
Bell, Alexander Graham, 185–186, 249 Carnap, Rudolf, 27, 41–42
Berkeley Museum of Vertebrate Zoology, Carnegie Mellon University, 111
2, 12 Carroll, Jack, 123n6
298 Index
Challenger (space shuttle), 92, 247, 250–251, Endangered Species Act, 160, 165, 173
254, 255, 257 Enlightenment ideologies, 140, 142–144,
Chapman, David, 112–113 147
Chen, Donna T., 243 Enron Corporation, 92–93
Chomsky, Noam, 28 Epstein, Steven, 13
Civilization IV, 290 European Union, 286
climate change, 126 Evans, Robert, 2, 5–6, 10, 77, 286
Coburn, Cynthia, 193–194 Everglades, 127, 129–132, 151
cognitive diaries, 95 exchange relations, anthropology of, 34–35
cognitive science, 1, 136 expertise, types of, 55–59, 77
Collins, Harry, 2, 5–6, 10, 21n3, 47, 66,
77, 114, 286 fabric, environmentally friendly, 13–14
Columbia (space shuttle), 243, 246, 253–257, Federal Reserve System, 240
262 Female Health Company (FHC), 244, 281,
complex adaptive systems, 127, 146–147, 283–287
149, 265–266, 268–271, 275 Feynman, Richard, 92, 255
computing Finin, Tim, 110–111
biology, 110–111, 115–122 Fisher, Erik, 73, 209–219, 222, 224–227, 292
combi-chem, 112–114 Food and Drug Administration, 271–272,
technology, 11 282
Consortium for Science Policy and Foucault, Michel, 42
Outcomes, 2 Franklin, Alan, 229n2
contributory expertise, 55–57, 59, 115 Fugelsang, J. A., 289
Cooper, Cynthia, 260–261 Fuller, Boyd, 50
corporations. See business world fundamentalism, 139, 141, 143
cybercommunities, 1
G. D. Searle, 281
Decision Theater, Arizona State University, 90 Galison, Peter, 2, 5, 7–8, 12, 26–27, 29–30,
Dialogues between Two Cultures (seminar), 41, 80, 89, 121, 189, 245, 266, 292
214–215, 221, 226 Gargiulo, William, 281–282
distributed cognitive systems, 185–189, 190 Gates Foundation, 268, 286
Drell, Sidney, 32–33 Gioia, Dennis, 250–252
Dunbar, Kevin, 289 Gleevec (imatinib mesylate), 271
Dutton, Tom, 28, 97n15 Gooding, David, 82, 293
Gorman, Michael E., 21n, 22n4, 71, 80–82,
earth systems engineering and management 88, 122n, 185–187, 214, 243, 267, 270
(ESEM), 99n31, 144–145, 148–150, 152 Grameen Bank, Bangladesh, 85
Easterly, William Russell, 94, 98n22 gravitational wave detection, 14
Ebbers, Bernie, 258, 261 Gray, Elisha, 185, 249
Edison, Thomas A., 185 Griesemer, James, 3, 12, 46
Eichenwald, Kurt, 92 Grinnell, Joseph, 2–3
Einstein, Albert, 30–31 Groves, James, 81, 267
Elhai, Jeff, 119–120 Gudeman, Stephen, 34
Index 299
Ham, Linda, 254, 256 Jobs, Steve, 73, 237

Harris, Joe, 39 Johnson, Mark, 80
Hayles, N. Katherine, 137 journalists, embedded, 216, 223
Heine, Heinrich, 134 Journal of Psychology of Science and
Hessel, Lasse, 282 Technology, 289
HIV/AIDS, 10, 13, 243–244, 272, 275, 281,
284–285, 287 Kaiser, David, 43
HIV/AIDS Coordination Programme Katrina (hurricane), 131
(Zimbabwe), 284 Kazakhstan, 128
humanist, embedded, 215–216, 223–228 Kelly, Walt, 295
humanistic engineering, paradigm conflicts Kepler, Johannes, 249
in, 213, 218–219, 224 Klahr, David, 111
Husbands, Rick, 255 KnowOS, 117
Hutchins, Edwin, 187–189, 191, 202 Kreutz, Dave, 78, 96n10
hydrological engineering, cultural Kuhn, Thomas S., 1, 7, 26–27, 36, 42, 79, 93,
significance of, 131–132 97n15, 188–189, 246–248, 250, 277
Kusch, Martin, 21n3, 114
IBM, 76, 77, 96n11 Kyoto Protocol, 126
Imitation Game, 60–68
Inca Empire, conquest of, 190 Langley, Pat, 111
information technology, 1, 90–91, 136–138, Latour, Bruno, 46–47, 183–184
140–142 Law, John, 183–184
interactional expertise, 2, 5, 12–16, 47–48, Leeper, Mary Ann, 244, 281–283, 285–286
58–59, 68–69, 84, 158–159, 168–169, Lego, 76
175–177, 245, 294–295. See also Lewis, Clayton, 215, 222
Imitation Game Lexar, 90
defined, 53 Licklider, J. C., 88
within Periodic Table of Expertises, 55–59 Logo, 115
interlanguages, 8, 10, 13–14, 20, 27–29, Lorenz, Konrad, 238
41–43, 80, 97n15, 235, 245, 281. See also Los Alamos National Laboratory, 87–88, 292
out-talk Lund, Robert, 251
computer tools as, 89
creole, 8, 83 Mackenzie, D., 98n25
diagrams and symbols, 43–45, 72–73 Mahajan, Roop L., 73, 209–219, 222, 224–226
jargon, 8 Malinowski, Bronislaw, 238
and metaphors, 81–82 Manhattan Project, 87–88, 268
pidgin, 8, 28 Marxism, 50, 127, 131–132, 140–141
Internet-based business, 236, 239 Massachusetts Institute of Technology
Iraq, 128 Radiation Laboratory (RadLab), 255
Isabella, Lynn, 243 Mauss, Marcel, 35
MDL, Inc., 114
Jenkins, Lekelia D., 72, 77, 238, 292, 293 Mehalik, Matthew M., 22n4, 72, 88
Jeremijenko, Natalie, 214 Mendel, Gregor, 45
300 Index
mental models, 15, 72, 82, 85–91, 130, 141, Oetinger, Bolko von, 73–74
144–145, 147, 181, 185–192, 198, 200, Oppenheimer, J. Robert, 87–88, 134, 291
202, 211, 227–228, 231, 233, 236, 248–252, Oracle, 77
257, 262 organizational deviance, normalization of,
Mills, Ann E., 243 243, 246–247, 249–250, 257, 261
Minkowski, Hermann, 27 out-talk, 29, 33, 47–48. See also interlanguages
moral imagination, 246, 251–252, 255
Mortenson, Greg, 94 Pakistan, 94
Morton Thiokol, 251, 256 Papert, Seymour, 115
Mulloy, Lawrence, 251 Parrish, O. B., 281–282
music industry, 236–237, 239 Periodic Table of Expertises, 55–59
Plsek, P., 265
n3, 39–40, 42–43 Poincaré, Henri, 31–32
Nanosim, 290–291 Pound, Robert, 25
nanotechnology, 1, 17–18, 33–34, 47, 49, Powell, Elizabeth, 244
80–81, 136, 216–217 Precautionary Principle, 149
National Aeronautics and Space President’s Emergency Plan for AIDS
Administration (NASA), 92, 251, Relief (PEPFAR), 243–244, 268, 272–273,
254–257 275–276
National Endowment for the Humanities, protocol, conceptions of, 218–219
214
National Institute of Standards, 84 Quine, W. V. O., 27, 41–42
National Institutes of Health, 48
National Marine Fisheries Service (NMFS), railroads. See under technology systems
72, 160–161, 163–174 Ratzel, Friedrich, 234
National Oceanic and Atmospheric Reagan, Ronald, 253
Administration, 173 relativity theory, 30–31
National Science Foundation, 2, 48, 77, resource regimes, 127–132, 149, 151–152
80–81 robotics, 136
Converging Technologies, 137–138 Rocha, Rodney, 254, 256
Strategic Plan, 1 Roco, M. C., 138
NBRIC, 136–137 Rorty, Mary V., 243
Netherlands Bank, 286 Rosson, Mary Beth, 123n6
network states framework, 182, 190–193,
206–207 Schaffer, Simon, 46
Neurath, Otto, 27, 41 Schilhab, Theresa, 67
New Orleans, 131 school district policy alignment, 181,
no-man’s-land, 234 193–194
non-government agencies, 286–287, accountability, 193, 204–205
290 systems mapping, 194–206
Norman, Donald, 11, 86 Schumpeter, Joseph, 232
Normand, Troy, 259–260 Schunn, Chris, 82
Nyamukapa, Daisy, 284 Schwinger, Julian, 37–39
Index 301
science and technology studies (STS), technology, as cultural process, 131–132,

54–55, 79 135–136, 138–139
Seabright, Paul, 99n28 technology systems, 127, 132–133
Sea Grant, 160, 163–164, 167–171 and economics, 132–134, 137
Sea Turtle Recovery Team, 172 and human biology, 137–138
Securities and Exchange Commission, 261 railroads, 133–136
Sen, Amartya, 81, 85 Tesco, 236
Senate Committee on Commerce, Science, Texas Shrimp Association, 161, 173
and Transportation, 1 Thompson, E. P., 25
Senge, Peter, 248, 252 Thomson, William, Lord Kelvin,
service science, management, and 45–46
engineering (SSME), 71, 75–78, 80, 83, Thorpe, Charles, 98n24
85–86, 88, 91, 93–95, 99n31, 100n33 time
service scientists, education and training of, absolute, 30–32
79–80, 94–95 uniform systems of, 133, 136
service systems, 75–76 Tomanaga, Sin-Itiro, 37–38
Shapin, Steven, 98n24 Toshiba, 90
Sherif, Carolyn, 253 trading zones
Sherif, Muzafer, 252–253 classification of, 9, 21, 83–86
Shrager, Jeff, 71–72, 81, 89, 292 Cold War, 48–50
Siegler, Bob, 111 as complex adaptive systems, 266
Sigma Six, 240 and cultural stability, 29
Simmel, Georg, 234 defined, 8, 32, 245, 281
Simon, Herb, 111 enforced, 9–10, 158, 163, 173, 175
Societal Dimensions of Nanotechnology, 290 evolution of, 16–20, 21, 157, 176–177,
Socio-Technical Integration Research project, 222–224, 240
229n3 fractionated, 11–14, 68, 73, 158–159,
software engineers, as mediators of science, 165, 175, 226, 228
107–108 interlanguage, 8, 9, 11, 158–159, 228
Southeastern Fisheries Association, 161 limitations of concept, 44–47, 143–144
Spillane, J., 194 and neoliberal reductivism, 34
Spinardi, G., 98n25 normative dimensions of, 245
Spohrer, Jim, 71 and power asymmetry, 37, 39
Stacey, Ralph, 269–270 and self-identity, 241
Star, Susan Leigh, 3, 12, 46 subversive, 10–11, 158–159
Stevinus, Simon, 28 suspended, 226–228
studies of expertise and experience (SEE), Trading Zones, Interactional Expertise,
54–59, 95, 289 and Interdisciplinary Collaboration
Stupka, John, 261 (workshop), 2, 232, 289
Sullivan, Scott, 259–260 translational medicine, 267–268
Traweek, Sharon, 35
Taton, Arnaud, 119 T-shaped expertise, 78
Taussig, Michael, 35 Turing, Alan, 60
302 Index
turtle excluder device (TED), 72, 157–158,

163–178
TED Voluntary Use Committee, 170–172
Twenty-first Century Nanotechnology
Research and Development Act (NRDA),
216–217, 219
Uhl-Bien, M., 265

United Nations Food and Agriculture
Organization, 128
University of Colorado, Boulder, 209–210.
See also Fisher, Erik; Mahajan, Roop L.
University of Pittsburgh. See school district
policy alignment
University of Tokyo, 98n26
Uzbekistan, 128
Vaughan, D., 246–247, 254

Veblen, Thorstein, 35
Vinson, Betty L., 259–260
Wallace, David, 98n26

war, 38
water management, 131
Wear, Larry, 247
Webster, Daniel, 134
Werhane, Patricia H., 81, 243–244, 270
Whitman, Walt, 135
Wikipedia, 233
Winokur, Mark, 215, 221, 226
Wisconsin Pharmacal, 281–283
Wittgenstein, Ludwig, 28
Wizard, 111–112
Woolgar, Steve, 184
World Bank, 286
WorldCom, 243, 257–262
Yavapai Indians, 13
Zajonc, Arthur, 214

GORMAN - 2010 - Trading Zones Interactional Expertise

Uploaded by

Copyright:

Available Formats

GORMAN - 2010 - Trading Zones Interactional Expertise

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

GORMAN - 2010 - Trading Zones Interactional Expertise

Uploaded by

Copyright:

Available Formats

SCIENCE, TECHNOLOGY, AND SOCIETY

TRADING ZONES AND INTERACTIONAL EXPERTISE

Cross-disciplinary collaboration increasingly characterizes today’s science and engineering research.

THE MIT PRESS

Janet Abbate, Inventing the Internet

Charles Bazerman, The Languages of Edison’s Light

Pablo J. Boczkowski, Digitizing the News: Innovation in Online Newspapers

Geoffrey C. Bowker, Memory Practices in the Sciences

Louis L. Bucciarelli, Designing Engineers

H. M. Collins, Artificial Experts: Social Knowledge and Intelligent Machines

Kristen Haring, Ham Radio’s Technical Culture

Christine Hine, Systematics as Cyberscience: Computers, Change, and Continuity in Science

Anique Hommels, Unbuilding Cities: Obduracy in Urban Sociotechnical Change

Donald MacKenzie, Inventing Accuracy: A Historical Sociology of Nuclear Missile Guidance

Donald MacKenzie, Knowing Machines: Essays on Technical Change

Donald MacKenzie, Mechanizing Proof: Computing, Risk, and Trust

Helga Nowotny, Insatiable Curiosity: Innovation in a Fragile Future

Richard Rottenburg, Far-Fetched Facts: A Parable of Development Aid

Charis Thompson, Making Parents: The Ontological Choreography of Reproductive Technology

Dominique Vinck, editor, Everyday Engineering: An Ethnography of Design and Innovation

Creating New Kinds of Collaboration

edited by Michael E. Gorman

The MIT Press

For information about special quantity discounts, please email special_sales@mitpress.mit.edu

Library of Congress Cataloging-in-Publication Data

1 Introduction: Trading Zones, Interactional Expertise, and Collaboration 1

2 Trading Zones and Interactional Expertise 7

3 Trading with the Enemy 25

4 Interactional Expertise and the Imitation Game 53

II Applying Trading Zones and Interactional Expertise to Domains of Practice 71

5 Service Science: A New Expertise for Managing Sociotechnical Systems 75

6 From Wizards to Trading Zones: Crossing the Chasm of Computers in

7 Authenticity, Earth Systems Engineering and Management, and the

9 A Network States Approach for Mapping System Changes 181

10 Embedding the Humanities in Engineering: Art, Dialogue, and a

III Ethics and Trading Zones 243

12 Using Trading Zones to Prevent Normalized Deviance in Organizations 245

13 Viewing Trading Zones Developed to Advance Health as Complex

14 Creating Trading Zones across Continents and Economies: The Female

15 Conclusion: Future Research on Trading Zones and Interactional Expertise 289

Recent National Science Foundation initiatives like nanotechnology emphasize col-

Kuhn, T. S. 1962. The Structure of ScientificRrevolutions. Chicago: University of Chicago Press.

Harry Collins, Robert Evans, and Michael E. Gorman

Trading Zones as the Locus of Incommensurability

Interlanguage Trading Zones

A General Model of Trading Zones

Enforced Trading Zones

Subversive Trading Zones

Fractionated Trading Zones

language alone, it cannot be too heavily stressed, interactional expertise is tacit-

Trading Locations and the Fractal Approach

The Evolution of Trading Zones

INTERLANGUAGE COLLABORATION FRACTIONATED

SUBVERSIVE COERCION ENFORCED

becomes a subfield of chemistry or another existing disciplinary culture. In other

Funding Trading Zones

1. The key reference for this is Galison 1997.

action irrespective of the intentions. “Polimorphic” actions, in contrast, need to be understood

6. For discussion of this and other cases, see Scott 1998.