Technological Development and Science in The Industrial Age

TECHNOLOGICAL DEVELOPMENT AND SCIENCE
IN THE INDUSTRIAL AGE

BOSTON STUDIES IN THE PHILOSOPHY OF SCIENCE
Editor
ROBERT S. COHEN, Boston University
Editorial Advisory Board
THOMAS F. GUCK, Boston University

ADOLF GRÜNBAUM, University 0/Pittsburgh
SAHOTRA SARKAR, Boston University
SYLVAN S. SCHWEBER, Brandeis University
JOHN J. STACHEL, Boston University
MARX W. WARTOFSKY, Baruch College 0/
the City University ofNew York
VOLUME 144
TECHNOLOGICAL
DEVELOPMENT
AND SCIENCE IN THE
INDUSTRIAL AGE
New Perspectives on the
Science-Technology Relationship
Edited by
PETER KROES
and
MARTIJN BAKKER
Department 0/ Philosophy and Social Sciences,
Eindhoven University 0/ Technology, The Netherlands
Springer-Science+Business Media, B.V.

Library of Congress Cataloging-in-Publication Data
Teehnologleal develop.ent and selenee In the Industrlal Age : new

perspeetlves on the selenee-teehnology relationship / edlted by
Peter Kroes and Martljn Bakker.
p. e• . -- (Boston studles In the phl10sophy of se lenee ; v.
144)
Ineludes blbllographleal references and Index.
ISBN 0-7923-1898-6 (aeld-free paper)
1. Technology--Congresses . 2. Sclenee--Congresses. I. Kroes,
Pete r , 1950- U. Bakker , Mart 1jn, 1956- IU . Ser 1es .
0174.B67 vol. 144
[T6]
600--de20 92-1611
ISBN 978-90-481-4186-9 ISBN 978-94-015-8010-6 (eBook)

DOI 10.1007/978-94-015-8010-6
Printed on acid-free paper
All Rights Reserved

© 1992Springer Science+Business Media Dordrecht
Originally published by Kluwer Academic Publishers in 1992.
Softcover reprint of the hardcover 1st edition 1992
No part of the material protected by this copyright notice may be reproduced or
utilized in any form or by any means, electronic or mechanical,
including photocopying, recording or by any information storage and
retr ieval system, without written perm ission from the copyri ght owner.
T ABLE OF CONTENTS
PREFACE vii
PETER KROES and MARTIJN BAKKER I Introduction:

Technological Development and Science
WALTER G. VINCENTI I Engineering Knowledge, Type of

Design, and Level of Hierarchy: Further Thoughts about
What Engineers Know... 17
EDWIN T. LAYTON Jr. I Escape from the Jail of Shape;

Dimensionality and Engineering Science 35
PETER KROES I On the Role of Design in Engineering

Theories; Pambour's Theory of the Steam Engine 69
ANDRIES SARLEMIJN and MARC DE VRIES I

The Piecemeal Rationality of Application-Oriented Research .
An Analysis 01 the R&D-History Leading to the Invention 01 the
Plumbicon in the Philips Research Laboratories 99
ROBERT FOX and ANNA GUAGNINI I Life in the Slow
Lane : Research and Electrical Engineering in Britain, France,
and Italy, ca. 1900 133
HARRY LINTSEN, GIEL VAN HOOFF and GEERT
VERBONG I Mechanical Engineering in The Netherlands in the
Nineteenth Century: Technology without a Professional
Community 155
EDA KRANAKIS I Hybrid Careers and the Interaction of
Science and Technology 177
JOHN M. STAUDENMAIER, S.J . I Science and Technology:

Who Gets a Say? 205
ARIE RIP I Science and Technology as Dancing Partners 231
NAME INDEX 271
v
PREFACE
This book is the offspring of a conference on the relationship between

science and technology held at Eindhoven (The Netherlands) in
November 1990. The editing of the proceedings of a conference always
requires a concerted effort by all authors involved. It is the task of the
editors to direct their work to a common goal. They are, however,
dependent on the authors. In this respect, the editors of this volume have
been extremely lucky. In preparing these proceedings we experienced a
strong commitment to th is project by the authors. We take this
opportunity to express our thanks for their cooperation; it was a pleasure
to work together. Furthermore we would like to thank Stephen H.
Cutcliffe and Steven L. Goldman for their comments on an earlier version
of the manuscript. Finally, our gratitude goes to Robert S. Cohen, Editor
of the Boston Studies in the Philosophy 01 Science and to Annie Kuipers
of Kluwer Academic Publishers for their support and advice.
PETER KROES
M A RT IJ N BAKKER
vii
P. KROES AND M. BAKKER
INTRODUCTION:
TECHNOLOGICAL DEVELOPMENT AND SCIENCE
Almost two decades ago, in March 1973, the famous Burndy Library
conference on The Interaction 01 Science and Technology in the
Industrial Age took place. 1 One of the explicit aims of this conference was
to bring together historians of science and technology to discuss the
mann er in which science and technology influence each other. In his
Foreword to the proceedings of the conference, Robert E. Kohler
remarked that a better insight into this matter required " a breaking down
of barriers between the two groups of historians" .2 Indeed, an adequate
study of a topic such as the interaction between science and technology
requires a trespassing of disciplinary boundaries, or even better the
abolition of those boundaries . Kohler already saw a hopeful sign that
something like that was happening: at universities historians of science
and technology were increasingly placed in one and the same department.
According to Kohler, this institutional development would insure that
"the history of science will cease once and for all to be the history of
isolated pieces of intellectual software, or the history of technology to be
the history of isolated pieces of hardware". 3 Twenty years later, the
cooperation between historians of science and technology seems indeed to
have intensified, but the disciplinary boundaries are still operative."
The basic issue at the conference was the relationship between
scientific knowledge and technological devices. Most participants agreed
on the inadequacy of the "technology is applied science" point of view,
that is of models postulating a linear, sequential path from scientific
knowledge to technological invention and innovation. In this way, science
is considered to be the " prime mover" of technology. From such models
it is but a small step (by adding the widely accepted postulate of an
internal developmentallogic for science) to some form of technological
determinism. Instead the "relative autonomy" of technology and techno-
logical development with regard to science was generally acknowledged
and stressed.
Apparently, semantic issues concerning the meaning of notions like
"science" , "technology", "engineering" and "applied science" heavily
dominated the Burndy Library conference .' The problem of distinguish-
P. Kroes and M . Bakker (eds.), Technological Development and Science in the Industrial
Age, 1-15 .
© 1992 Kluwer Academic Publishers.
2 PETER KROES AND MARTIJN BAKKER
ing science from technology received especially close attention; it was

generally concluded that it was not possible to draw a sharp distinction
between the two ." There was also a clear recognition of the fact that, in
practice, any distinction between science and technology may be strongly
influenced by ideological elements .' Against the background of policy
issues concerning the funding of "pure" and "applied" science and
ongoing discussions about projects like Hindsight (1966) and Traces
(1968), these topics were of course of primary importance."
The interest in the science-technology issue at the Burndy Library
conference was itself not an incidental affair. In his recent review of
trends in the history of technology for the American Historical Review,
Staudenmaier observes that the topic of the relationship between science
and technology has attracted a lot of attention in Technology and Culture
ever since the first appearance, in the late 1950s, of this by now most
prestigious journal in the field ." Despite all the effort devoted to this
topic, progress in this field has, however, not been very impressive.
Staudenmaier complains that the same kind of questions (and the same
kind of answers , one would like to add) are posed over and over again. In
Technology's Storytellers he reaches the conclusion that historians of
science and technology "have not developed a thematic language about
science and technology that they find appropriate for the history of
technology. Instead we find evidence of the disintegration of the theme
despite the fact that it addresses issues that are clearly of great interest in
the field" . 10 This disintegration shows itself in a lack of consensus about
the definitions of the basic notions involved ("science" and " tech-
nology") and in a whole array of conflicting theories, models and
approaches to the issue.
Progress in this field has indeed been severely hampered by a persistent
demarcation problem; the identification and isolation of the two principal
elements, science and technology, remains a formidable obstacle. Clear
and unambiguous classification criteria are still lacking; what is called
"science" and "scientific" in one context is often referred to as
"technology" and "technological" in another, and vice versa. Even if it is
admitted that the notions "science" and "technology" do not represent
homogeneous classes of activities, and that there may be a cons iderable
overlap in these activities, the demarcation problem remains, unless, of
course, the rat her radical conclusion is accepted that any distinction
between science and technology makes no sense.
In the twentieth century, professional and institutional borderlines are
INTRODUCTION 3
somehow not reliable indicators for science and technology. Nowadays, it

is no exception that mission oriented research conducted in industrial
research laboratories by researchers trained as engineers engenders
important scientific breakthroughs or that scientists working at
universities or academic centers come up with important technological
inventions. Scientists and engineers migrate from one context to the other
as if no borderline exists." At the cognitive level, the situation is also
obscure; clear identification marks for scientific and technological
knowledge (for instance, for engineering theories and scientific theories)
are still absent. In view of their intimate relationship, the metaphor of
science and technology as siamese twins'? looks very appropriate; science
and technology seem to form an organic whole, from which it is
impossible to carve out science and technology as separate entities without
brutally mutilating both.
One way out of this situation would be, indeed, to give up, in the
context of the study of technological development, the distinction
between science and technology altogether. At the Burndy Library
conference this was, for instance, suggested by Arnold Thackray. 13 In our
days, this attitude is rather popular in social constructivist circles.:" A
look at recent work in the history and philosophy of technology, however,
shows that it is still pervaded by the distinction between science and
technology, no matter how controversial this distinction may be.
Somehow "science and technology" talk is an integral part of the
conceptual background of historians and philosophers of technology (and
science). Apparently, it cannot easily be avoided.
The distinction between science and technology is deeply rooted in
western culture. It was canonized in Greek antiquity by Plato and
Aristotle, whose works in so many respects still influence western
thinking. Part of the current problems with the science-technology
dichotomy rnay be due to the fact that Greek ideas about the nature of
science and technology still bias our thinking, whereas in fact these ideas
may be outdated, particularly by the birth of modern science in the 16th
and 17th centuries. It appears that philosophers and historians of science
and technology have failed to develop an adequate substitute for or at
least adapt the Greek distinction in the light of later developments. It falls
outside the scope of this introduction to go into this matter in any detail.
The following considerations give some indication why a fundamental
reinterpretation of the notions "science " and " technology" is highly
desirable and necessary.
4 PETER KROES AND MARTlJN BAKKER
Within the Aristotelian philosophy, science in the strict sense (that is

"theoretical knowledge" which comprises "first philosophy", natural
science and mathematics) is concerned with knowledge for the sake of its
own". Scientific knowledge is knowledge of first principles and causes
from which other insights may be derived; science has a demonstrative
character. According to Aristotle, a complete scientific explanation of a
phenomenon, moreover, always consists of specifying the four causes
involved." Scientific knowledge consists of eternal, necessary truths
because the object of science is restricted to reality insofar as it is
unchangeable. The way by which man can reach scientific truths is
contemplation (8fWQHV), which is a faculty of the human soul.
Technology (techne), on the other hand, concerns the making of things
(1rOULV) . This, of course, also involves knowledge, called productive
knowledge by Aristotle; it differs from theoretical knowledge in that its
goal lies outside the acting subject and that its object is reality in sofar it
is changeable. Productive knowledge primarily concerns rules describing
how things can be made. Generally speaking, theoretical knowledge, in
particular physical knowledge, is not relevant for technology. The main
reason for this is that physical knowledge concerns nature as it exists,
independent of human intervention (theoretical knowledge is based on
contemplation) whereas human intervention lies at the basis of
technology. For Aristotle, science and technology clearly belong to two
different spheres of human experience (contemplation versus productive
action). Like many other Greek thinkers, his teacher Plato among them,
he considered science to be a higher form of human activity than
technology.
The rise of modern science in the 16th and 17th centuries made the
Aristotelian distinction between science and technology problematic, to
say the least. Several radical changes took place with regard to science.
First of all, experimentation became one of the cornerstones of modern
science. Instead of being a passive spectator, the modern scientist activeiy
intervenes in the course of natural phenomena in his laboratory. The
study of nature under artificial conditions is accepted as a valid way to
obtain knowledge about nature. Human intervention has become a
common element of science and technology; it no longer differentiates
science from technology as in the Aristotelian framework. Secondly,
mathematics became the principal tool for studying nature. Newton's
mathematical analysis of mechanics in his Philosophiae Naturalis
Principia Mathematica became a paradigm for the study of all physical
INTRODUCTION 5
phenomena, and it seems for all other sciences. It is interesting to point

out here that mathematics did not play any role in Aristotelian physics,
whereas it was employed in antiquity in some of the arts, for example
astronomy and mechanics." Finally, in connection with the introduction
of mathematics into science, the Aristotelian scheme for a scientific
explanation in terms of the four causes was abandoned. A mathematical
description of physical phenomena, such as motion and gravitation,
became more important than an analysis of the causes of these
phenomena. Quantitative functional relationships increasingly replaced
qualitative causal relationships. The not ion of causa finalis, which is so
prominently present in Aristotelian science, simply disappeared from the
new science.
The overall effect of these developments is that modern science and
technology have much more in common than in Greek antiquity and stand
in a much closer relationship to each other. Modern science, as opposed
to Aristotelian science, is intrinsically and not just incidentally"
technologically relevant because it is based on human intervention and
thus control of nature and, vice versa, science depends on technology for
the equipment needed for its experimentation. Modern science and
technology have become interdependent, primarily through the adoption
of the experimental method by modern science.
These changes are so far-reaching that the Aristotelian distinction
between science and technology loses its foundation and is no longer
appropriate for characterizing the new situation. Nevertheless, present-
day discussions of the nature of science and technology are still strongly
dominated by ideas stemming from the Aristotelian tradition. Time and
again, it is stated that science studies reality for its own sake, the aim of
science being true knowledge about the eternallaws governing the world,
whereas in technology knowledge itself is not the aim but a means to an
ulterior goal, that is, the design, construction and production of artifacts;
that science tries to explain phenomena by disclosing their true causes
whereas technology aims primarily at the control of phenomena for
human ends; that science studies natural phenomena and technology
artifacts created by human action." Such characterizations of science and
technology show the strong influence of the Aristotelian heritage; they
have become part of the common-sense conception of science and
technology, with all the dangers inherent in common-sense conceptions.
The reason for this situation may be that a viable alternative to the
Aristotelian conception of science and technology is not available .
It is rather paradoxical to observe that, precisely in the era in which the

rapprochement between science and technology star ted to bear fruit for
the development of both science and technology - i.e. the second half of
the 19th century - a strong tendency arose, especially in Western Europe,
to view science and technology as two different activities . It is the age of
what has been called the emancipation of technology and its associated
rhetoric of "applied/practical" or "engineering" science on the one hand
and of the rhetoric of "pure" science and the pursuit of truth for its own
sake, free of all kinds of constraints on the other. Engineers, for their
part, portrayed scientists as eccentric people out of contact with reality,
studying problems of no practical relevance. Scientists, by contrast,
although they stressed that mankind could only benefit from the
advancement of science, nevertheless considered themselves elevated
above the technological exploitation of scientific knowledge; that was
considered to be the "dirty" business of engineers. As regards the rhetoric
about pure science, it is as if Plato or Aristotle had become alive to
reiterate the Greek ideal of science!
By the end of the 19th century the professional and institutional
separation of science and technology was more or less a fact in many
countries. There were different educational systems for engineers and
scientists and they were organised in different societies. Ideological
commitments not only strongly influenced public images of science,
technology and their relationship, but also directl y affected research
policy and the organisation of research. Instead of working together,
engineers and scientists at times opposed each other.
A nice illustration can be found in Thomas Hughes ' American
Genesis. He discusses the establishment of the Naval Consulting Board
during the Great War. 20 This Board was mainly composed of inventors
and engineers. Representatives of the American Physical Society and of
the National Academy of Seiences were deliberately excluded from the
board, which was headed by Edison, who is characterized by Hughes as
one of the greatest "independent inventors" of that time. The scientists,
eager to show that they were able to render service to the state, especially
in time of war, responded with the establishment of the National Research
CounciI. These two institutions competed in developing submarine
detection devices after Germany had resumed submarine warfare in 1917.
According to Hughes: "Antisubmarine research and development
precipitated not only cooperation but competition among inventors,
engineers, and academic and industrial scientists. Each group believed
INTRODUCTION 7
that it had unique qualities to bring to the task, and each made claims on
national resources to increase its numbers and activities". 21 Apparently,
engineers and scientists considered it more opportune to emphasize their
own, specific abilities and resources for solving a given problem than to
point out their common and complementary means for dealing with it.
What we see here could be characterized as a quarrel between engineers
and scientists about whether the detection of submarines is a "tech-
nological" or "scientific" affair. 22 The interests at stake, of course, are
very high, for the answer to this question has far-reaching consequences
for the funding of research in this field. But the simple fact that science
and technology had become competitors for the solution of a certain
problem illustrates the degree to which modern science and technology
had converged since the emergence of modern science. It is also illustrated
by the fact that, around the turn of the century, industry recognised the
technological impact of science and, undisturbed by the rhetoric of
engineers and scientists, started to lure scientists into its laboratories.
The Greek idea - that science and technology are essentially distinct
activities - was not only part of the 19th century rhetoric; it still is very
influential today, and has not been replaced or adapted. Time and again
this idea is endorsed by engineers and scientists themselves, more on
ideological grounds, it seems, than on the basis of a careful and critical
analysis of what engineers and scientists do. But the historiography of
science and technology has also contributed to the continuation of this
idea. Although many historians of technology admit that the distinction
is problematic, it is nevertheless an integral part of most of the work done
in this field. This, of course, is a rather dangerous situation, for it may
lead to an uncritical adoption of the self-image of scientists and engineers
and its associated rhetoric.
As an example of how implicit assumptions with regard to the science-
technology dichotomy may influence the historiography of technology, it
is worthwhile to pause for a moment at the notion of a "Second Industrial
Revolution". This notion is increasingly becoming part of the standard
vocabulary of historians of technology and is usually associated with a
revolutionary phase in the development of technology around the turn of
the century due to the (systematic) technologieal exploitation of scientific
knowledge. It refers to the emergence of a science-based technology and
is related to the creation of industrial research laboratories and the
employment of scientists in industry. Ernst Homburg has carefully
analysed the origins of this notion and how it was introduced in the
history of technology." The notion was first used in the fifties of this
century, when it referred to the technological revolut ion due to
automation and nuclear technology and its social impact. It was employed
in a "politic-programmatic"" way to argue for or against social changes
related to the ongoing technological revolution. In the beginning of the
sixties, it was transformed into a historical concept by Geoffrey
Barraclough and David Landes. According to Homburg the original
concept of a Second Industrial Revolution was based upon two
presuppositions which fitted very well with the ideology of the fifties with
regard to the relation between science and technology, viz. the idea that it
is possible to draw a coherent distinction between science and technology
and the linear science-technology-innovation model. He argues that in
the transformation of the notion of a "Second Industrial Revolution"
into a historical concept part of the ideological connotation and its
presuppositions were carried along without being subjected to a critical
examination. That is why, in his opinion, the use of this notion is
problematic, especially because criticism of the last decades on the
conception of science and technology as monolithic activities and on the
linear science-technology-innovation model has not resulted in a
reinterpretation of this notion.
Gur intention is not to argue that, in view of the above considerations,
any distinction between science and technology is obsolete and that
therefore quest ions about how science and technology influence each
other are senseless. We agree with Homburg that " fruitful historical
research into the relation between science and technology is certainly
possible and also desirable" ,15 on condition that new conceptions of
science and technology are developed . We have merely tried to indicate
that research in this field appears to be guided by an outdated conception
of science. In order to pose the right kind of questions it is necessary to
abandon the Greek distinction between science and technology, and to
search for areinterpretation of these notions such that a more fruitful
framework arises for analyzing the interaction between science and
technology. Such areinterpretation has, in our opinion, to be based
primarily on an analysis of what engineers and scientist actually do
instead of on what they claim they do.
A new promising approach to the old science-technology issue may
already be developing. Staudenmaier's analysis of TC's papers addressing
the science-technology issue has led hirn to the conclusion that in these
papers historians of technology are not so much dealing with the science-
INTRODUCTION 9
technology relationship than with the nature of technological knowl-

edge." As areaction to the "technology is applied science" thesis, the
analysis of characteristic features of technological knowledge emerges as
a significant new topic. For Staudenmaier this topic - Characteristics 0/
Technological Know/edge - constitutes a potential new theme for
approaching the long-standing problem of the interaction of science and
technology." In his opinion, technological knowledge is a special kind of
knowledge, distinct from scientific knowledge, which derives its " unique
cognitive qualities from the tension between technical design and its
ambience, which defines the nature of technology itself" . 28 He
dist inguishes and discusses four particular characteristics, namely:
scientific concepts, problematic data, engineering theory and technical
skill . Aseries of excellent studies into the nature of technological
knowledge by Walter Vincenti in Technology and Cu/ture illustrates this
new approach."
Note that Staudenmaier relates the unique cognitive status of
technological knowledge to the "tension between technical design and its
ambience" . This is a crucial remark . In the first place, the nature of
technological knowledge is related to the notion of design, a notion which
plays no role in scientific knowledge. Secondly, it means that the specific
nature of technological knowledge and the way it develops can only be
understood in the framework of the tension between an artefact and its
context and not, as the technology is applied science model implies,
against the background of the development of scientific theories. More
generally, this point of view entails that the analysis of the development of
technological knowledge has to be based on a thoroughly contextual
approach. In other words, the contextual style has to be the main
methodological style in analyzing the nature of technological knowledge
just as it is in the current history of technology .30
The analysis of the cognitive status of technological knowledge not
only calls for a historical but also for an epistemological approach. In
other words, the philosophy of technology becomes involved as weIl as the
history of technology. Although Staudenmaier does not draw this
conclusion, his suggestion implies that it may be fruitful to remove the
existing fence between the history and philosophy of technology. The
study of the unique cognitive properties of technological knowledge
requires a common effort on the part of the historians and philosophers
of technology. When analyzing the nature of technological knowledge ,
historians of technology may benefit from epistemological studies of
characteristic features of technological knowledge (as opposed to

scientific knowledge), whereas such epistemological studies, in order to be
relevant to an understanding of the actual historical development of
technology, have to be based in turn on extensive historical research.
Let us turn briefly to the philosophy of technology" to see what it has
to offer in this respect. The philosophy of technology as aseparate
philosophical discipline is now about one century old. For a long time it
has been dominated by a metaphysicallethical tradition in which
questions about the nature of technology in the context of human
experience were the central focus of interest. The works of scholars like
Dessauer, Heidegger and Jaspers belong to this tradition. The kind of
problems posed and the methods of analyzing them generally have no
direct bearing on the issues discussed here .
Of greater interest for the problems under consideration is the recent
emergence of a new current in the philosophy of technology , which is
strongly focused on the cognitive dimension of technology; more and
more epistemological studies of technological knowledge started to
appear after about 1970.3 2 The anthology Contributions to a Philosophy
of Technology , with the subtitle Studies in the Structure of Thinking in
the Technological Sciences, published by Friedrich Rapp in 1974, was a
first attempt to establish an analytical philosophy of technology, which
was to be focused on "a methodological and even an epistemological
analysis of the theoretical structure and the specific methods of procedure
characteristic of modern technology" .3 3 This kind of philosophy of
technology was considered to be a counterpart to the philosophy of
science. The papers collected by Rapp dealt, among other topics, with the
relation between natural sciences and technology, technology as applied
science, methodological differences between the engineering and natural
sciences, classification schemes for the engineering sciences, and different
aspects of the design process .
Although the book contains interesting studies of various aspects of
technology (including its relation to science), it is nevertheless not the kind
of philosophy of technology which may be expected to make important
contributions to the science-technology issue discussed above. The reason
for this is that no attention is given to the history of technology. Rapp
states in the Introduction that "the articles contained in this book
generally do not take into account the historical development of
technology" .34 The various topics are analysed from a systematic (logicall
methodological), a-historical point of view. As a result, preconceived
INTRODUCTION 11
notions of technology and science play an important role, whereas these

notions would have to be grounded in historical research .
In order to contribute to the study of the interaction between science
and technology, the philosophy of technology has to take a historical
turn ." According to Elisabeth Ströker the development of a "historical
philosophy of technology" is necessary for a fruitful systematic analysis
of technology since "history - and history alone - provide[s] all those
concepts that form part of the repertoire of the philosophical analysis of
technology " . 3 6 Indeed, this is the way to avoid the pitfall of forcing the
analysis of technology into the mould of preconceived notions. She argues
that the conceptual framework for analyzing technology has to be
acquired through historical reconstructions of its subject area; only in this
way "can the philosophy of technology effectively arrive at the
development of its systematics" . 3 7
Staudenmaier's and Ströker's remarks point in the same direction: the
convergence of the history and philosophy of technology as a necessary
step for an adequate study of the interaction between science and
technology . New perspectives on the science-technology issue may be
disclosed in this way. The cognitive dimension of technology has to be
approached in a historical-philosophical way. Let us immediately add a
warning here. Any attempt to interprete this proposal from either side in
an " imperialistic" way would turn it into a sterile exercise. The philos-
ophy of technology is not to be subordinated to the history of technology
by reducing or dissolving philosophical problems to historical ones; nor is
the reverse to be attempted. J ust as science and technology themselves,
they form a siamese pair.
It is against this background that a group of historians and
philosophers of technology at the University of Technology of Eindhoven
decided to organize a conference on Technological Development and
Science in the 19th and 20th Centuries," Although the title may suggest
so, it was not to be arepetition of the Burndy Library conference. The
idea was to bring together historians and philosophers of technology . The
central topic was to be the influence of science on the development of
technology during the last two centuries and special attention was to be
given to the theme of science and the nature of technological knowledge,
as weil as the role of science in engineering education and the engineering
profession. In recent years historians and philosophers of technology
have started to ask new questions about the science-technology issue.
They are questioning the arguments on which the distinction is based in
the hope of demystifying the distinction and of finding new, fruitful

starting points for research in this area . The conference was intended as a
platform for historians and philosophers to exchange views on these
matters.
The papers published here cover various aspects of the science-
technology issue. First comes aseries of papers dealing with cognitive
matters. Vincenti discusses the use and generation of engineering
knowledge in what he calls normal and radieal design, and Layton
analyses the role of scale models in technology and the emergence of
dimensional analysis. Kroes argues that, at the theoretical level,
engineering knowledge distinguishes itself from scientifie knowledge by
the prominent place of design and design parameters . Sarlemijn and De
Vries introduce the notion of piecemeal rationality of application-
oriented research , as opposed to striet rationality and serendipity. The
thread running through all these papers is the emphasis on the notion of
design. Issues of a historiographie nature are then addressed. Fox and
Guagnini question the assumption, whieh is widespread among historians
of technology, that a strong commitment to research is essential for
technologieal and economic progress , and Lintsen et a/. argue that, in The
Netherlands, the industrialisation process came about without the
involvement of professional communities . In the last three papers, the
authors broaden the perspective on the science-technology issue by
adding also social aspects . Kranakis puts forward the proposal to use
" hybrid careers " as a key notion for analyzing the interaction between
science and technology. Staudenmaier discusses the role of science and
technology in the context of creating controlled environments and of the
emergence of new ideological perspectives. In the final paper, Rip
pro poses a new approach to the science-technology issue, whieh is based
on the conception of science and technology as search processes and
discusses qualitative transformations of the science-technology complex .
The papers presented at the conference and the discussions made it
c1ear that the problem of the interaction between science and technology
was still as alive as ever. A major methodologieal shift in dealing with the
issue, though, appears to have occurred over the past few decades.
Whereas at the Burndy Library conference the demarcation problem, and
with it semantie issues about the notions "science" and "technology" and
the "technology is applied science" thesis, were at the foreground of the
discussions, these topies, aIthough addressed occasionally, were not
dominant at the Eindhoven conference. Instead of general, "global"
INTRODUCTION 13
interpretation schemes and models of the interaction between science and

technology, detailed empirical case studies of cognitive and institutional
connections between "science" and "technology" constituted the hard
core of the conference. Given the situation that more and more of these
"loeal" case studies are becoming available, one of the main challenges to
historians and philosophers of technology is to search for recurrent
patterns and at the same time to develop appropriate conceptual
frameworks for describing those patterns. This bottom-up approach may
eventually disclose new (global) perspectives on the role of science in
technological development .
March 1992
The Editors
NOTES
1 The proceedings were published by Nathan Reingold and Arthur Molella in a special issue
of Technology and Culture; see Reingold and Molella (1976).

2 Reingold and Molella (1976), p. 623.
J lbid.
4 There are, for instance, no major international journals devoted to the history of science
and technology. A superficial bibliographical search resulted in the following list of journals
explicitly devoted to the history of science and technology : Historical Studies in lrish Science
and Technology, Centaurus, Scientia Canadensis and the Dutch Gewina .
s See Reingold and Molella (1976), p. 625.
• See in particular Otto Mayr's contribution to the proceedings .
7 According to Kohler the whole issue is basically ideological; Reingold and Molella (1976),
p.621.
• For more information about Hindsight and Traces, see Rosenberg [1982), pp. 207-216 .
9 Staudenmaier (1990), p. 718.
11 See Kranakis's contribution to this volume .
12 See A. Sarlemijn, ' Science and Technology - Present Relations in Historical Perspective',
in Sarlemijn and Kroes (1990), pp . 3-21.

IJ Reingold and Molella (1976), p. 645.
'4 They do not give up the distinction between science and technology altogether, but reduce
it to a purely social construction. See, for instance , T.J . Pinch and W.E . Bijker, 'The Social
Construction of Facts and Artifacts: Or How the Sociology of Science and the Sociology of
Technology Might Benefit Each Other', in Bijker (1987), pp , 17-50, and Bijker (1990).
" For the following, see Aristotle, Nicomachean Ethics, 1139b-1I4I b ; l177 s_l177 b ;
Metaphysics, 982b ; 1025b-1026 s; Eudemian Ethics, 1216b •
I. These four causes are : the material. formal final and efficient cause.
14 PETER KROES AND MARTlJN BAKKER
17 In antiquity, mechan ics was considered to be an art; it was transformed into a science in
the 16th and 17th centuries.

18 See Aristotle, Eudemian Eth ics, 1216
b15.
" For a critical discussion of the opposition between artificial and natural
phenomena/objects with regard to science, see F. Rapp, 'Technology and Natural Science
- a Methodological Investigation', in Rapp (1974), pp. 93-114; see also E. Str öker, 'Philos-
ophy of Technology: Problems of a Philosophical Discipline', in Durbin and Rapp (1983),
pp . 323-336.
20 See Hughes (1989), p. 118 ff.
21 Hughe s (1989), p. 123.
22 According to the British physicist Ernest Rutherford submarine detection was " a problem
of physics pure and simple" (quoted in Hughes (1989), p. 123).

23 Homburg (1986), pp . 367-385 .
24 See Homburg (1986), p. 373.
" Homburg (1986), p . 368; the translation is ours ,

2. In his review article Staudenmaier (1990), p. 718, qualifies Vincenti 's work as " the last
genuine methodological innovation on the topic" .

3 0 Staudenmaier (1985), p . 84.
31 For a survey of the philosophy of technology, see C. Mitcharn , 'Philosophy of
Technology' in Durbin (1980), pp . 282-363 .

3 2 For instance: Mitcham and Mackey [1972; especially part I); Lenk and Moser (1973);
Rapp (1974); Huning (1974); Rapp (1981).

33 Rapp (1974), p. vii.
3 4 Rapp (1974), p. xii.
3 S Ju st as the philosophy of science did in the sixties and seventies for analyzing the way
science develops and progresses.

3 6 Elisabeth Ströker, 'Philosophy of Technology: Problems of a Philosophical Discipline',
in Durbin and Rapp (1983), p , 333.

37 Ströker in Durbin and Rapp (1983), p . 334.
38 The conference was held on 6-9 November 1990 at Eindhoven University of Technology
and attracted considerable attention; about seventy scholars attended the conference,
coming from The Netherlands (26), the United Kingdom (13), the USA (9), Germany (9), the
Scandinavian countries (6), France (4), Spain (3), Israel (I) and Australia (I). Nine "invited"
papers (published in this volume) and fort y "contributed" papers were presented . Most of
the participants (about 50) were historians of technology; it was in fact no suprise that the
philosophers of technology were in the minority, since the history of technology is much
better institutionalised as a discipline than the philosophy of technology.
INTRODUCTION 15
REFERENCES
Bijker, W.E. et al. (eds.): 1987, The Social Construction of Technological Systems,
Cambridge Mass .: MIT Press .
Bijker , W .E .: 1990, 'Do Not Despair : There Is Life after Constructivism', Kennis en
Methode, No . 4, pp . 324-345 .
Durbin, P .T . (ed.) : 1980, A Guide to the Culture of Science, Technology, and Medicine,
New York : Free Press .
Durbin, P .T . and Rapp, F. (eds.): 1983, Philosophy and Technology, Dordrecht : Reidel.
Homburg, E.: 1986, 'De 'Tweede Industriele Revolutie', Een problematisch historisch
concept', Theoretische Geschiedenis 13(3), pp . 367-385 .
Hughes , T .P .: 1989, American Genesis, Penguin Books .
Huning, A.: 1974, Das Schaffen des Ingenieurs; Beiträge zu einer Philosophie der Technik,
Düsseldorf: VDI -Verlag .
Lenk , H . and Moser, S. (eds.) : 1973, Techne, Technik, Technologie, Pullach bei München:
Verlag Dokumentation .
Mitcham, C. and Mackey, R. (eds.): 1972, Philosophy and Technology, New
York /London: The Free Press .
Rapp, F. (ed.): 1974, Contributions to a Philosophy of Technology, Dordrecht: Reidel.
Rapp , F.: 1981, Analytical Philosophy of Technology, Dordrecht: Reidel.
Reingold, N. and Molella, A. (eds.): 1976, 'The Interaction of Science and Technology in
the 1ndustrial Age' , Technology and Culture 17(4), pp . 621-742.
Rosenberg , N.: 1982, Inside the Black Box: Technology and Econom ics, Cambridge:
Cambridge UP .
Sarlemijn, A. and Kroes , P . (eds .): 1990, Between Science and Technology, Amst erdam :
North-Holland.
Staudenmaier, J .M .: 1985, Technology's Story tellers, Cambridge Mass .: MIT Press .
Staudenmaier, J .M .: 1990, 'Recent Trends in the History of Technology', American
H istoricol Review 95, pp. 715-725.
WALTER G . VINCENTI
ENGINEERING KNOWLEDGE, TYPE OF DESIGN,

AND LEVEL OF HIERARCHY : FURTHER THOUGHTS
ABOUT WHAT ENGINEERS KNOW . . .
This article comes at a point in my work that is both advantageous and

awkward. The awkwardness comes from the fact that I have recently
published a book under the title What Engineers Know and How They
Know It (hence the allusion in the title of the present piece).' This book
contains most of what I think I know about what engineers know, and
what I offer here will not be essentially new. The advantages arise because,
like most authors, I have been having second thoughts about what I have
written and about ideas I think I see more clearly now. I shall attempt here
to repackage and summarize those ideas in a way that - I hope - will make
more explicit the historiographie and epistemological structure behind
them. This structure did not appear so clearly when I was occupied with the
nuts and bolts of the work . A diagram has also occurred to me that
embodies some of the key ideas in an easily remembered and suggestive
form . I will present and discuss it in the concluding part of this material.
PRELIMINARY REMARKS
To the extent that scholars concern themselves with technological

knowledge, they typically do so in the context of new devices and unusual
events. Most historical studies, I think it fair to say, focus on invention,
innovation, and the activities that take place in research establishments.
Such preoccupation is understandable - novelty is dramatic, relatively
visible and accessible to study, and central to the important problems of
technical, economic, and social change. In history, as in everyday life,
newness and variety are always more exciting than day-to-day routine.
For the study of technological knowledge, however, such focus may be
going at the problem the hard way - the cognitive content of creativity is
notoriously difficult to specify. Moreover, something essential may be
overlooked in the process. General history has found it profitable in
recent decades to move from study of great individuals and unusual events
to examination of everyday life. Something analogous may be useful for
the historical study of technological knowledge .
17
P. Kroes and M . Bakker (eds.), Technological Development and Science in the Industrial
Age , 17-34.
18 WAL TER G. VINCENTI
Philosophical analysis of technological knowledge shows correspon-

ding bias. Most philosophers who have looked into the epistemology of
technology have come from the already established study of the
epistemology of science. Despite the best of intentions, they would be only
human to bring preconceived notions with them and seek to apply existing
concepts where they may be inappropriate or at best only part of the story .
An excellent recent book of articles in the field, for example, was titled
The Nature 01 Technological Knowledge. Are Models 01 Scientific
Change Relevant?" Since I have used the articles to advantage, I may be
ungracious to observe that the title could as weil have asked if such models
are "irrelevant". That it was phrased as it was may reflect something
about unconscious attitudes and assumptions. Suppose, contrary to fact,
that scholars had embarked on the study of technology before that of
science. In place of arguing about whether technology is applied science,
they might have been debating if science is not, at least in part, theoretical
technology.:' Would not the received wisdom about the relation between
science and technology then differ from what it is? That, of course, must
be conjecture. A danger surely exist, however , that preoccupation with
novelty on the one hand and undue influence from study of science on the
other could lead to a partial or faulty epistemology of technology.
My own work has attempted to examine the knowledge employed in
conventional everyday technological activity - what I call normal
technology - on more or less its own terms . In taking this direction, I was
doing only what came naturally from a lifetime of experience as an
engineer - the ideas above are the wisdom of hindsight. No influence was
possible from philosophy of science because I knew nothing about it at the
outset. The emphasis on normal technology was instinctive because my
career as a research engineer and teacher has been spent producing and
organizing the kind of knowledge such technology requires . Though some
of this knowledge may have been novel in itself, it had in the end to be
useful to the workaday engineers who constitute the vast majority of the
profession. I also knew that most of the students I instructed would join
that majority. The kind of knowledge I dealt with was conditioned
accordingly. The fact that an experienced engineer would focus,
automatically and at first unconsciously, on normal technology for study
of engineering knowledge may itself say something about the nature and
importance of such knowledge.
As the preceding sentence implies, I shall deal here only with engineer-
ing knowledge, not technological knowledge generally . Engineering I take
ENGINEERING KNOWLEDGE, DESIGN, AND HIERARCHY 19
to be defined for present purposes as the practice of organizing the design,

production, and operation of an artifact or process that transforms the
physical world to some recognized human end. The crucial word here is
"organizing" . It distinguishes engineers from other technologists, such
as, for example, draftspersons, shop personnei, and airplane pilots, who
carry out the tasks that engineers organize. Engineers, that is, constitute
a subcategory of technologists . By the same token, engineering
knowledge - the knowledge engineers use - does not make up the whole
of the knowledge required in technology. What I have to say is a
contribution to the epistemology of engineering, and hence at the same
time of technology; it does not claim, however, to address the entire scope
of technologieal knowledge.
To keep the task within bounds, I limit my work also mainly to design.
Design provides, in fact , a major - some say distinguishing - aspect of
engineering (though, curiously, it rarely appears as a subject heading in
the annual index to Technology and Cu/ture or the bibliography of
current publications that appears each year in that journal). As indieated
in my book, the epistemologieal ideas can , I think, be extended to
production and operation, though that task remains to be done .
As developed in the book, the ideas grow out of five historieal case
studies from my professional field of aeronauties. Four of these appeared
earlier, with minor differences, in Technology and Cu/ture; one of them
is new. Here I can only present the ideas themselves, without the historieal
narrative and discussion leading to them . I believe, however, they are
grounded in historical fact.
NORMAL AND RADICAL DESIGN
When we examine knowledge in the context of everyday design, ideas

emerge that seem to me fundamental for epistemological and historiogra-
phie concerns. One of these is the concept of normal, day-to-day
technology already mentioned . For me the notion arose as I examined
case studies chosen for other reasons or because, somehow, they seemed
the "right" thing to look at. I found later that I was in fact seeing what
Edward Constant, in his book The Origins 0/ the Turbojet Revolution ,
had already called "normal technology" . This activity - "what techno-
logieal communities usually do" - Constant defined as comprising "the
improvement of the accepted tradition or its application under 'new or
more stringent conditions''' . 4 The designers whose needs I was studying

can thus be described as doing normal design (my extension, not
Constant's), this being the design component of what engineers "usually
do" . Engineers carrying out such design know at the outset how the device
in question works, what it usually looks like, and that, if properly
designed along such lines, it has a good likelihood of accomplishing its
task. They learned these things at some point in their enculturation into
the design community; they most likely begin their designs without giving
them conscious thought. Though such knowledge had to be generated at
some time in the past, it is now simply taken for granted.
As recognized by its definition, normal design makes up the major part
of engineering enterprise. As remarked by a reviewer of a draft of my
book, "For every Kelly Johnson [a well-known designer of innovative
aircraft] there are thousands of useful and productive engineers designing
from combinations of off-the-shelf technologies that are then tested,
adjusted, and refined until they work satisfactorily" . Such activity
constitutes the majority of what goes on in the vast design offices of firms
like General Motors, Boeing, and Bechtel. Though little examined by
scholars, it would be remarkable if such a large and widespread activity
had no epistemological importance.
Two of the things that engineers take for granted in normal design
warrant names of their own. For "how the device ... works" I use Michael
Polanyi's term operational principle. In Polanyi's words, a device's
operational principle defines how its "characteristic parts ... fulfil their
special function in combining to an overall operation which achieves the
purpose" of the device." All devices (which I take to include processes and
static structures as weIl as machines) possess such a principle. For the
winged device called the "airplane" , for example, the operational
principle prescribes that the upward force needed to balance the vehicle's
weight be generated by propelling a rigid surface forward through the
resisting air . This principle, new when put forward by Sir George Cayley
in 1809, distinguishes the airplane from the helicopter, which obtains both
lift and propulsion from an engine-driven rotor, and from the
ornithopter, which attempts to do the same with flapping wings. All
airplane designers now take Cayley's principle for granted. They know
from over a century of experience that it works . Similar knowledge of an
operational principle underlies all normal design.
The second concept in normal design - what the device "usually looks
like" - I term the normal configuration . By normal configuration I mean
ENGINEERING KNOWLEDGE , DESIGN, AND HIERARCHY 21
the general shape and arrangement commonly agreed to best realize the
operational principle . It becomes arrived at and agreed upon (perhaps
implicitly) in the formative years of the device. Though less strictly
determinative than the operational principle, it too constitutes a given in
normal design. The normal configuration of the airplane over most of its
existence has been the engine-forward, tail-aft biplane (mainly) until the
1930s and monoplane thereafter. This arrangement was arrived at ,
primarily in France, before World War 1. Since then and until recently, it
has only rarely occurred to designers that an airplane could or should be
arranged differently.
The operational principle and normal configuration together define
the normal design of a device. They form the basis for the " accepted
tradition" that Constant spoke of in his characterization of normal
technology (though he did not allude to them specifically). What we may
call radical technology then involves change in either the device's accepted
configuration or its operational principle. In the latter case, the configur-
ation too must change - a new and appropriate normal configuration will
have to be arrived at once the new operational principle has been
established. Considerations like these underlie our ideas of invention and
innovation. The design component of radical technology I shall refer to as
radiea/ design.
Radical and normal design constitute, of course, two limits of a
spectrum more than they do a dichotomy. The degree of radicalness is
obviously greater when the operational principle is changed than when
only the normal configuration is altered. And either the principle or
configuration can be modified in some degree rat her than completely
replaced . As usual in these matters, the distinctions are sometimes
difficult to make . They do, however, provide a useful tool for analysis.
The operational principle and normal configuration afford clear
instances of engineering, as against scientific, knowledge. They may be
analyzed and in some cases even triggered by the findings of science; they
are in no way, however, contained in or dictated by those findings. As
stated by Polanyi, "The complete [by which he meant scientific]
knowledge of a machine as an object teils us nothing about it as a
machine ." The operational principle and normal configuration call for
added acts of insight and experiment, usually by inventors or engineers.
The foregoing statements accord with a perceptive distinction by
Herbert Simon in his The Scienees 01 the Artificial. 6 As stated by Simon,
the natural sciences deal with how things are. Engineering design, like all
design, deals with how things ought to be. The operational principle and
normal configuration of an airplane, for example, are in no sense
knowledge of how a flying machine innately is; they are knowledge of how
a particular kind of flying machine ought to be to serve certain purposes.
They provide instances par excel/ence of engineering (as distinct from
scientific) knowledge . As I point out in my book, all knowledge for
engineering design, including the necessary attendant knowledge of how
things are, serves finally to implement how things ought to be. That, in
fact, is the criterion for its usefulness and validity . (Similar remarks would
apply to engineering knowledge for production and operation.)
USE AND GENERATION OF KNOWLEDGE
Still another idea, obvious also from everyday design, dominates the
epistemology of engineering . In design activity (as also in production and
operation), the practical use to which knowledge is put is paramount. The
necessity that the knowledge be useful both motivates the knowledge and
determines its nature. When need for its use disappears - as when , say, the
reciprocating steam engine was replaced by the steam turbine - the
knowledge becomes neglected and, for practical purposes, forgotten. In
engineering knowledge, practical use is of the essence. This requirement is
obvious, but in the preoccupation with scientific knowledge it tends at
times to be forgotten .
This is not to say that that use is unimportant in science; it is important,
but in a very different way. Science I take to be a search for understanding
of observable phenomena, such understanding itself constituting a form
of knowledge. This search for understanding or knowledge is open-
ended . In science, understanding hence finds use in generating more
understanding - or, equivalently, knowledge finds use in generating more
knowledge. As put somewhat differently by historian Hugh Aitken,
"Most of the informational output of science - the new knowledge
generated - is channelled back into science itself. "7
In the foregoing paragraphs, knowledge figures in engineering activity
as a means to a utilitarian end , in scientific activity as a means to more
knowledge (and hence in a sense as an end in itself). This difference
conforms with Simon's distinction between science as dealing with how
things are and engineering design with how things ought to be. Engineers,
of course , also use knowledge to generate more knowledge. Such use,
ENGINEERING KNOWLEDGE, DESIGN , AND HIERARCHY 23
however, is less requisite than its use in designing artifacts. The essential
point for present purposes is this : By employing knowledge in design as
weIlas in generating more knowledge - that is, for two uses instead of one
- activity in engineering is distinctively asymmetrie from activity in
science.
This asymmetry has an epistemologieally important result. In science,
where knowledge is used to generate more knowledge, the institutional
loci of generation and use are overwhelmingly one and the same - the
scientifie research laboratory. In engineering , generation and use may
also take place together in the engineering research-and-development
laboratory. As numerous studies testify, however, distinguishing between
scientific and engineering knowledge proves difficult if one examines
them, as commonly done, in the context of these similar and overlapping
research institutions. In engineering, where practical application is
paramount, however, the situation is in fact different. There knowledge
finds its main and defining use in a separate institution - the industrial
design office. Distinction between the two kinds of knowledge thus
becomes operationally possible . To identify the nature of engineering
knowledge in terms of its use - which, by definition, should be the starting
point of any epistemology of engineering - we can look to see what
knowledge is employed in the design office. For scientifie knowledge we
do the same in the scientific laboratory. Failure to notice this possibility
may be one effect of approaching the history and epistemology of
engineering with biases from the history and epistemology of science. An
outstanding historian of technology, whose work I have used and admire,
despaired of distinguishing between engineers and scientists when he
perhaps fell into this trap by seeing both groups as working "in
laboratories of like appearance" (emphasis added) ." Focusing on normal
rat her than radieal technology helps avoid this pitfall.
To visualize the foregoing ideas, I find Fig. I helpful. The key notion
here is the differentiation between knowledge used by scientists and
engineers (solid-line boxes) and knowledge generated by those
communities (solid-line band) . At the level of use, distinction between
knowledge in science and engineering can be made as explained above;
representation by separate boxes is therefore realistic . At the level of
generation, as likewise observed, such distinction is problematic;
representation by a spectrum is thus more credible, with knowledge
generated by scientists toward the left and by engineers toward the right.
The activities that produce this knowledge then appear as a corresponding
24 WAL TER G . VINCENTI
- - ~1e-;;lIi1c k-;;O';Ied~- - 7/ 7/,7h'/ /7-0;- Engl"';rln; k;'wle,j;.- - -

generelIng sellv ilie. /j. - Comblned or %
generelIng seIlv ilIes
r- %
(cerrIed out by
%
selentls ts prlmer lly
Interpenet reUng
sellvnles~ .
%
(cerried oul by
englneers prlmerlly
- 1
I __ ~O~~~S~I~ _ r / L/"( ~<~'/L /"(i jT _ '~ ~e~ce~l~ty~ _
I I I I I I
t t t t t
I
I Knowledge genereted Knowledge genersled
I by selenllsls by engineer.
I
I
I
I Knowledge used by
selenI1sts
(to generete
Knowledge used by
englneers
(to design srtlfsets end,
more knowledge ) sec:ondarily, generate
more knowledge)
_____ t- _
I I
I I
I Design setlvlty I
I I
L I
Fig. 1. Knowledge and its generat ing activit ies,
spectrum (dashed-Iine band), with purely scientific activity far to the left
and purely engineering activity correspondingly to the right . The activities
of hybrid individuals or of groups working together in laboratories then
fall in the dashed area in the center. It is this spectrum of activities that
generates and determines the spectrum of knowledge directly below.
Scientists and engineers for their respective use then draw on this
generated knowledge as needed from anywhere along the spectrum. The
same piece of knowledge may thus be found identically (or appropriately
modified) in both solid-line boxes. If so needed, an item mayaiso show up
in one box but not in the other. As indicated in the diagram, the
knowledge used by engineers finds employment primarily in design
(dashed-Iine box) and secondarily in the knowledge-generating activities
of engineers. Knowledge used by scientists goes preponderantly toward
the single use of generating more knowledge. The essential asymmetry
between engineering and science thus appears clearly in the diagram. In
my book I attempt in effect to unpack the contents of the solid-line box

on the right. 9
The diagram, of course, should not be taken literally. It idealizes and
oversimplifies an obviously complex and problematic situation. I find it
useful, however, as a mnemonic device and a framework for thinking
about engineering knowledge. Though derived here from considerations
of normal design - since that's where it emerges most plainly - I see no
reason it should not apply (though with more flexibility) to radical design
as weIl.
DESIGN HIERARCHY
Another concept, basic to design though not reflected in the diagram, is

that of design hierarchy. Most complex modern "devices" exist in fact as
hierarchically organized systems of interrelated and mutually dependent
components and subcomponents. Design of the elements of this hierarchy
goes on more or less independently in separate groups and subgroups of
engineers. For a conventional airplane, for example, once the operational
requirements have been translated into concrete, quantitative specifi-
cations for the designer (a preliminary level known as project definition),
the design levels can be set down as follows. The example here is
deliberately simplified; different and more complex breakdowns are not
unusual.
1. Conceptual design -layout of general arrangement and proportions of
the airplane to meet the project definition.
2. Major-component design - division of overall project into fuselage
design, wing design, electrical-system design, landing-gear design, etc .
3. Subdivision of components from level 2 according to engineering
discipline, e.g., overall landing-gear design into mechanical,
structural, and aerodynamic landing-gear design . These disciplinary
areas will typically involve subsidiary design (or selection from
standard items where possible) of subcomponents, e.g., retraction
devices, shock absorbers, tires , wheel brakes, etc., within mechanical
design .
Such successive division resolves the airplane problem into smaller
sub problems accessible to semi-isolated solution. The overall design
process then takes place iteratively, both vertically and horizontally,
throughout the hierarchy.
Another point is essential: Whether design is radicalor normal is

independent of loeation in the hierarehy. Normal design ean (and usually
does) prevail throughout, but radical design ean arise at any level. A
radically eoneeived airplane eonfiguration, for example, eould
ineorporate a more or less normal landing gear made up of highly
eonventional (even eatalog) subeomponents. An otherwise wholly
eonventional airplane might utilize a shoek absorber designed on a
radieally new prineiple by an inventive subeomponent engineer. Radical
design, though, rarely happens simultaneously at all levels; if it had to,
very little invention and innovation of eomplex deviees would oeeur .
Mueh, perhaps most, design effort for even "new" hierarehically
organized deviees proeeeds on the basis of existing, off-the-shelf
knowledge. The epistemology of engineering needs to ineorporate this
faet.
As remarked earlier, historiography of teehnology has eoneerned itself
in the past mostly with radical teehnology. Sueh foeus , moreover, has
been mainly at the upper levels of design hierarehy. Aetivity at those levels
tends to be visible and loosely struetured and henee provoeative and
dramatic. It appeals in the same way as do the breakthrough events
typically examined in the historiography of scienee. My own studies, in
looking at normal design, do so at the lower hierarehieallevels. That, like
my foeus on normal teehnology, was at first instinetive and uneonseious.
EPISTEMOLOGICAL RESUL TS
Studying normal aetivity at the lower levels, I find, has its dividends.
Kinds of engineering knowledge present themselves that might otherwise
be overlooked. My intention here has been to clarify this underlying
historiographie and epistemological orientation of my work. I have spaee
here to mention only a few results, more to illustrate the nature of the
findings than to provide an adequate and eomplete exposition. The reader
may reeognize some of them from my published articles; they are
elaborated in my book . (Generalized eategories into which the examples
fall are italicized in the following numbered aeeount.)
(I) The operational principle and normal eonfiguration have already
been pointed to as basic concepts in normal design. (2) To help translate
these concepts into a concrete device, engineers devise criteria and
specifications; a sophisticated example here is the specifications used by
commercial and military customers to obtain airplanes with tlying

qualities satisfactory and congenial to pilots. (3) To go from such specifi-
cations to quantified drawings for the shop (a more protracted task than
this simple statement makes it sound!), designers commonly employ
theoretical tools; a typical instance here is the method of control-volume
analysis used in design of all sorts of fluid-flow devices. As in this case,
these tools are frequently based on knowledge from science; even so, the
knowledge must usually be extended and reformulated to make it
applicable to engineering problems. (4) Application of theoretical tools in
turn requires many kinds of quantitative data; as example here we can cite
the empirical parametric data on air-propeller performance used to select
the best propeller for a given airplane design in the 1920sand 1930s. (5) To
arrive at their designs, engineers also call upon a range of practical
considerations, such as the rule - part of most tlying-quality specifi-
cations - that piloted airplanes should be inherently stable
aerodynamically but not too much so. This knowledge, derived from
practical tlight experience over a span of years, provides an excellent
example of engineering dealing with what ought to be. (6) Finally,
designers need at their disposal various design instrumentalities -
procedures, ways of thinking, and judgmental skills - on which the
process of design depends; prominent among these are the procedures
designers use in their efforts to optimize a device for a given application.
As the category of design instrumentalities suggests, engineers need to
know both of what philosopher Gilbert Ryle calls "knowing how" and
"knowing that", that is, knowledge of how to carry out tasks as weil as
knowledge of fact ." Knowing how includes, for present purposes, not
only knowledge of how to carry out design, but also of how to obtain or
generate the particular knowledge used in that process . To generate the
propeller data mentioned above, for example, research engineers brought
to bear the time-honored method of experimental parameter variation,
where the parameters of the device and its operation are varied
systematically in suitably contrived tests and the performance repeatedly
measured. Since the tests had to be made at reduced scale, the
experimenters also had to know how to use theoreticallaws of similitude
to transfer results from the working scale model to the full-sized device.
The method of parameter variation is sometimes followed by scientists,
but for very different purposes; the method for dealing with scale rarely,
if ever, is needed in science. These examples constitute only a sampIe of
what turns up. Ample evidence exists that engineers do indeed, in Edwin
28 WAL TER G . VINCENTI
Layton's words, develop "their own bodies of more or less systematic

knowledge to meet the needs of practice . "11
These examples suggest a theoretical generalization about engineering
vis-ä-vis science. In all the cases examined, the criterion of validity for
knowledge in engineering can be phrased, Does it help in designing
something that works in solution of some practical problem? (To
designing we would need to add producing or operating to cover all
engineering.) The meaning of "works" in a given situation is subject to
interpretation, but the necessity for utility remains . The corresponding
criterion for knowledge in science can no doubt be argued
philosophically. I suggest it is more or less as folIows: Does it help in
understanding some particular features of the universe? What one means
by "understanding" and "universe" , are, of course, also subject to
interpretation. Whatever the wording or interpretation, however, the
essential distinction must lie between intellectual understanding, or
explanation, for science and practical utility for engineering. I do not
mean to imply that engineers have no concern with understanding in
relation to their designs or scientists with utility in respect to their
experimental apparatus or even their theories. When the chips are down,
however, and choices must be made, the priorities of the communities
become clear. Engineers, when necessary, will put the desire for
understanding aside to get their job done on time; they may even use
questionable theories and assumptions if nothing else is available and
experience shows that they "work" . Scientists who chose to do so will
devote careers to the search for understanding that has no discernable
application; questionable information in science must be discarded or
unrelenting efforts made to resolve the questions .
A number of observations come to mind about the foregoing criteria. I
limit myself here to pointing out the parallel with the asymmetry discussed
earlier and exhibited in Fig. 1. As recognized in the criteria , knowledge is
a means to an end in both science and engineering, understanding in the
former , solution of some practical problem in the latter. "In science,
however [if I may quote from my book], the means acts directly to the end;
in engineering it acts through the intermediary of the "something" , usually
a material artifact, that is the immediate object of design (or production or
operation)." This is essentially the asymmetry in use of knowledge pointed
out earlier, expressed rather differently. As before, the asymmetry
conforms with Simon's distinction between science as dealing with how
things are and engineering design with how things ought to be.
So much, then, for the attempt to summarize the historiographie and

epistemologieal structure of my work. As I look back at what I have
written, it strikes me I may have packed in a heavy load for one paper to
carry . Such temptation, however, reflects a virtue of examining the needs
of normal design at the lower levels of hierarchy - it provides a great deal
to think about. It leaves little doubt, I think, of what historians of
technology have come to understand increasingly in recent years:
engineering knowledge warrants recognition as an epistemologieal species
in its own right.
ANAL YTICAL DIAGRAM
I come now to the promised diagram. As will be obvious, this addition

depends on and grows out of the concerns outlined above. It seems to me,
however, to have implications beyond the epistemology of technology. I
present it here in that light.
The diagram requires an additional pair of observations. These have to
do with the degree of technical constraint within which the designer works
at different levels of design hierarchy and at different locations in the
spectrum of design type (i.e., from radieal to normal). They will be
obvious to anyone experienced in design and are, I believe, reasonable in
themselves. First, at a given level in the hierarchy of design, constraint
tends to increase as the design type changes from radieal to normal.
Second, the general level of constraint at whieh this increase occurs tends
itself to increase as the focus of attention moves to lower levels in the
design hierarchy.
These ideas are perhaps best comprehended with aid of the diagram.
Here (see Fig. 2) type of design and level of hierarchy appear as
orthogonal independent variables in the horizontal base plane of a three-
dimensional space . Such representation embodies the earlier observation
that whether design is normal or radieal is independent of where it lies in
the level of hierarchy. If the degree of technieal constraint is plotted
vertieally, the tendencies pointed out in the preceding paragraph can then
be represented schematieally by a curved, three-dirnensional surface as
shown. The two tendencies appear as folIows:
At any given level of hierarchy, the degree of technieal constraint - the
elevation of the surface above the base plane - rises as the design varies
from radieal to normal. To make this reasonable, consider , for example,
Fig. 2. Degree of technical constraint as a function of type of design

and level of hierarch y.
a designer doing conceptual design of a complete airplane, that is,

working at the upper levels of design hierarchy. Such a designer
attempting a radical configuration or principle of operation obviously
works under less technical constraint (i.e., with more scope for
imagination) than someone doing anormal , conventional airplane. Such
difference follows almost by definition from the characterization of
normal design. The rise of the surface as one moves to the left at a fixed
level of hierarchy accords with this fact.
The general elevation at which this rise occurs also increases as the level
of hierarchy of concern goes down. In the airplane, for example, once
provisional decisions in conceptual design have set the weight, landing
speed, and arrangement of the vehicle - whether of radicalor normal
design - purely technical requirements of the landing problem take over
and greatly constrain the landing-gear designer's freedom to manoeuvre.
Landing-gear layout decisions at this intermediate level then put even
greater constraint on design or choice of shock absorbers and other
landing-gear subcomponents at still lower levels. When we combine this
effect of hierarchy with the effect of design type observed above, the
result is a degree-of-constraint surface that rises progressively toward the

rear of the diagram. A surface of this sort will exist for any complex
device. Though the comparison is one between apples and oranges,
downward movement in the design hierarchy seems to me likely to cause
relatively greater rise, in some sense, than change from radical toward
normal design. The diagram, as drawn, reflects this (perhaps debatable)
assumption.
As this last point suggests, the diagram must not be taken literally. The
variables - both independent and dependent - are intractable to exact
definition or ordinal ranking . The elevation of the constraint surface
represents only average of a range of likelihoods, not a precise value . The
diagram should be looked upon rather as a visual memory aid for a set of
ideas and relationships. As with all my discussion, it is a framework for
thinking and analysis , not a rigid structure.
To use the diagram, it must also be understood that design at different
levels of hierarchy in a given airplane does not all fall at the same location
between radical and normal. For the earlier example of the radical
configuration with a more or less normal landing gear of highly
conventional subcomponents, conceptual design of the airplane might be
visualized on the surface at Al, layout of the landing gear at A 2, and
design or choice of the subcomponents at A 3 • For the otherwise con-
ventional airplane with a radical shock absorber, conceptual design of the
airplane might He at B. and layout of the landing gear at B2 ; design of the
shock absorber would then appear at B3 • (I suspect the diagram is valid
in showing the radical shock absorber, though less constrained than the
conventional one, to be in some sense more so than the normal airplane
or landing gear. Such relative valuation, however, is only suggestive at
best.)
The diagram of Fig. 2 has implications, I suspect, beyond the
epistemological concerns from which it derives. As scholars have come to
appreciate, engineering design, like all of engineering, is a human
enterprise deeply embedded in a context of social needs and activities.
Such contextual social factors obviously exert great, sometimes
determining, influence on engineering outcomes. Such influence,
however - and here's where the diagram comes in - is far from uniform.
In engineering design, in particular, the non-uniforrnity would appear to
have direct relationship to the degree of technical constraint.
Such relationship is certainly theoretically plausible. As suggested
above, engineering in its broadest sense is part of a vast sociotechnical
system. The potential for social influence - even social determination or,
to use recent scholarly parlance, "social construction" - clearly exists at
all times. Technical constraints, however, have their base at least partly in
physical laws and requirements. To circumvent them is frequently
impracticable and in some cases physically impossible . When social
influences and technical constraints conflict, the constraints must often
prevaiI. Social influence would thus be expected to appear least where
elevation of our surface is highest, that is, in normal design at the
lowermost levels of hierarchy. It may be expected most at the lowest
elevations, to wit, in radical design at the upper hierarchical levels. I
submit that such theoretically argued outcome is by and large the fact.
As a case in point, consider again an aircraft and its landing-gear. At
the levels of project definition and conceptual design - that is, at the top
of the hierarchy - social factors obviously have wide scope for influence .
Even in normal design, the performance, size, and general layout of the
vehicle follow more or less directIy from the needs of the sometimes
contending social constituencies it serves. In radical design, where the
basic configuration or even principle of operation may be up for grabs ,
the connection is even closer. Across the full design-type spectrum, the
predominating social purpose and influence often leave their imprint in
the very appearance of the aircraft - a carrier-based fighter-bomber
visibly serves a different social purpose from a commercial four-engine
transport.
At the levels of component and subcomponent design, a different
situation obtains for both radical and normal design. As indicated earlier,
the problems of landing-gear layout and (especially) subcomponent
design are primarily technical problems amenable to and demanding
mainly technical solutions. The internal logic of the technical situation
takes over and leaves little room for direct social influence. The landing
gear and its subcomponents exhibit a great deal ofthe associated technical
purpose but very littIe, if any, social. (I do not regard the bare
requirement to get an airplane safelyon and off the ground as a
contextual influence, though it's obviously a social given.) The same
kinds of remarks can be made about other components and sub-
components of the airplane and, indeed, with regard to all complex,
hierarchically structured devices. Overall, the situation described in these
paragraphs reflects a crucial fact: At the upper levels of hierarchy, the
object of design functions typically as a component of a sociotechnical
system; at the lower levels it constitutes a component of a technical
ENGINEERING KNOWLEDGE , DESIGN , AND HIERARCHY 33
system, An airplane is as much a social construct as it is a technical; its

landing gear and landing-gear subcomponents are not.
I do not mean to imply in any way that engineering design is not a social
activity at all levels. Argument and negotiation - tradeoffs, power
struggles, personal agreements and disagreements, and so forth - take
place within and between the groups involved at all places in our diagram.
The greater the degree of technical constraint, however, the less chance
there is for this indispensable social interplay to affect substantially the
nature of the engineering outcome. At upper levels of hierarchy, where
the necessities of a sociotechnical system are commonly involved, the
interplay will frequently leave an imprint, especially in radical design. At
the lower levels, where problems are more purely technical, tradeoffs tend
also to be technical, and decisions get made on essentially technical
grounds. (Letting social factors take over at these levels may even be
courting disaster .) The foregoing does not mean that lower-level solutions
need be unique - different design groups attacking the same problem may
come up with different designs based on different judgments regarding
the technical tradeoffs. Even a radically different design may sometimes
appear. Whatever the solution, however, it embodies mainly technical
considerations and does not often reflect the social nature of the
arguments and negotiations that led to it. In some cases, the results are
even much alike - landing gears on airplanes of a given type appear very
similar unless one examines the mechanisms with an expert eye. At these
lower levels of hierarchy, technical factors set engineering problems and
control engineering outcomes. At upper levels, social factors often
(perhaps usually?) play the decisive role. The plausibility of the thesis for
the social construction oftechnology, as advanced, for example, by Pinch
and Bijker, 12 thus depends very much on the level of hierarchy (as weil as
type of design) under consideration.
The foregoing, of course, are generalizations that need further
examination and , perhaps, qualification. The diagram also has
implications, I believe - regarding technology policy, for example - that
I haven't time to pursue. Here I would like to add only a few brief
comments . The studies I have been engaged in have focused on
engineering knowledge and activity near the rear corner of the diagram.
The majority of work by the analytically-minded scholarly community, I
think it reasonable to say, has looked at people, events, and knowledge in
essentially the vicinity of the forward corner. Valuable though many of
the findings have been, the result has been to overlook points that show
up naturally here. Perhaps we would do weIl to expand our scope. Being

alert to the entire range of the diagram would at least help avoid the snare
(which ladmit to having succumbed to myself on occasion) of looking at
a limited area and generalizing that "Technology [implying all of
technology] is like this" or "Technological knowledge is Iike that" . Care
in defining, locating, and bounding the system or component under
examination in a given study - somewhat as engineers find essential in
thermodynamics - might be useful. Scholarship about technology has
reached the point, I believe, where more careful distinctions about what
we are looking at in a given case are necessary. An anthropologist
acquaintance who studies the computer industry teIls me he finds the ideas
embodied in the diagram helpful in sorting out what he sees. I offer them
here with the hope they may find use for others.
Stanford University
NOTES
I W.G. Vincenti, What Engineers Know and How They Know It: Analytical Stud ies from
Aeronautical History (Baltimore : Johns Hopk ins University Press, 1990).

2 R. Laudan (ed.), The Nature 01 Technological Knowledge. Are Models 01 Scientific
Change Relevant? (Dordrecht: D. Reidel, 1984).

J This thought comes from Robert McGinn .
4 E.W. Constant, The Origins 01 the Turbojet Revolution (Baltimore: Johns Hopk ins
University Press, 1980), p. 10. Constant's concept of normal technology is analogous to

(and derivative from) Kuhn's well-known concept of normal science; T.S . Kuhn, The
Structure 01 Scientific Revolutions (Chieago: University of Chicago Press, 1962).
s M. Po1anyi, Personal Knowledge (Chieago: University of Chicago Press , 1962), p. 328.
6 H.A. Simon, The Seiences 01 the Artificial, 2nd. ed. (Carnbridge, Mass.: MIT Press,
1981), pp . 132-133.
7 H.G.J. Aitken, Syntony and Spark - The Origins 01 Radio (New York: Wiley
Interscience, 1976), p. 314.
• O . Mayr, 'The Science-Technology Relationship as a Historiographie Problem',
Technology and Culture 17(October, 1976), pp. 663-673, quotation from p, 677.
• Fuller discussion of a truncated version of this diagram appears in the book.
10 G. Ryle, The Concept 01 Mind (London: Hutchinson, 1949), pp . 27-32 .
11 E.T. Layton , Review of O. Mayr, (ed.), Philosophers and Machines, Technologyand
Culture 18 (January, 1977), pp. 89-91, quotation from p. 89.

12 T.l . Pinch and W.E . Bijker, "The Social Construction of Facts and Artifacts: Or How the
SocioJogyofTechnology Might Benefit Each Other' , in W.E . Bijker, T.P . Hughes , and T.J.
Pinch (eds.), The Social Construction 01 Technological Systems (Cambridge , Mass.: MIT
Press, 1989), pp, 17-50 .
EDWIN T. LA YTON JR .
ESCAPE FROM THE lAlL OF SHAPE;

DIMENSIONALITY AND ENGINEERING SCIENCE
In V. Sackville-West's Poem, "The Land" the poet deals with craftsmen.

The poem reads in part:
All craftsmen share a knowledge. They have held
Reality down fluttering to a bench;
Cut wood to their own purposes; compelled
The growth 01 pattern with the patient shuttle,
Drained acres to a trench.
Control is theirs. They have ignored the subtle
Release 01 spirit from the jail 01 shape.
They have been concerned with prison, not escape.'
In trying to understand the rise of sciences associated with engineering
("engineering sciences" or "technical sciences"), we do not often turn to
poetry for inspiration. However, Victoria SackviIIe-West's suggestion
that there is a subtle essence of technology imprisoned in a jaiI of shape is
more than just a poetic conceit. We can think of technology as being
embedded in material artifacts. Attempts to und erstand engineering and
technology encountered obstacles which were rooted in the fact that
artifacts are three-dimensional bodies which have material properties.
Though shape is only one such property, it is particularly indicative in
historical perspective.
The efforts of engineers and others to wrest scientific knowledge from
artifacts embodied in brute matter has been one of the guiding themes in
the history of engineering. Three-dimensional objects cannot be increased
or decreased in size without changing shape or other material properties.
To cite a weIl-know example, the surface area of a sphere increases by the
square of the radius, the volume by the cube. Since Iiving organisms can be
roughly approximated by a sphere, this means that the rate of heat transfer
will vary with changes in scale . Heat dissipation is usually proportional to
the surface area, but metabolism, which generates heat, is usually a func-
tion of volume. Thus large animals (such as elephants) have problems
dissipating internaIly-generated heat, while small ones (such as humming-
birds) have problems preventing the dissipation of too much heat.
35
P. Kroes and M . Bakker (eds.), Technolo gical Development and Seience in the lndustrial
Age, 35-68.
36 EDWIN T . LAYTON JR .
Since man-made, three-dimensional artifacts have material properties,

and among these properties are extension, then artifacts have both shape
and size. Thus artifacts are influenced by scale in ways analogous to those
of geometrie bodies and animals. Some of the effects of scale changes are
fairly obvious. We are not surprised to discover that machines or
structures whieh are influenced by internally-generated heat are affected
by the changing ratio of surface area to volume in much the same manner
as animals.
Similarly, in pressure vessels, such as tubular boilers, the strength,
other things being equal, was inversely proportional to the size . That is, if
they are made of the same thiekness of the same material (say a partieular
grade of steel), pressure vessels with smaller diameters will be stronger
than those with larger diameters." The case of the boiler tube is similar to
the ratio of volume to surface area of a sphere. In both cases the change
with scale is a simple linear one, varying with the diameter in both cases .
Other changes with scale are less obvious. For example strength in a
solid may be influenced by the shape and material composition of the
crystals whieh are its building blocks. In many cases properties whieh were
thought to be purely mechanieal have turned out to depend upon shape.
For example, it was long assumed that the speed of a turbine's rotor could
be determined by a straightforward application of mechanics to the
problem. It turned out, however, that this speed is a function of both
shape and size (or scale). Each family of geometrically-similar turbines
has different speed behavior; to deal with this situation engineers define
a characteristie speed indieated by an index; this index is called the
"specific speed", whieh was originally defined as the rotations per minute
of a turbine wheel of unit diameter under unit head. It teIls designers how
rotor speed will vary with scale for a given family of geometrieally-similar
turbines . A turbine rotor one foot in diameter under one foot of head can
vary enormously in speed, depending upon shape, from as little as three
or four revolutions per minute (for impulse turbines of the Pelton type) to
more than two hundred (in the case of propeller turbines of the Kaplan
type) . Rotor speed is one of the most critieal factors in the rational design
of turbines and other hydraulie machines.:'
Simple linear relations involved in scale changes are relatively easy to
discover . But in other cases the relation of change of property or behavior
with change of scale is not linear. Indeed can be quite complicated,so that
the change in property with change in scale may be rather unobvious . For
exam ple the specific speed of a turbine varies inversely with the 5/4th
DIMENSIONALITY AND ENGINEERING SCIENCE 37
power ofthe head (or fall ofwater) and the square root ofthe horsepower.
For a centrifugal pump the specific speed varies inversely as the 3/4th
power of the head and the square root of the volume of flow.4 It is a
commonplace observation that seemingly minor changes in the shape of
an airfoil will change its performance (e.g. lift and drag) , and (less
obviously) that this performance is influenced by the shape of the flow
patterns in the surrounding fluid."
The change of observed behavior or properties of an artifact (e.g.
speed, strength, drag) with size are often called scale effects. Scale effects
were among the most important barriers to the development of a scientific
understanding of engineering artifacts. For most artifacts, engineers had
to understand elusive, often complex, dimensional relationships in order
to conduct model experiments or to develop accurate mathematical
theories. The struggle to und erstand these dimensional barriers to
scientific understanding played an important role in the historical
evolution of engineering. After centuries of struggles with elusive
dimensional problems, engineers and scientists by the 20th century were
led to the discovery (by the methods of "dimensional analysis") that
particular problems in engineering have an underlying dimensional
structure; once this structure was understood it enabled engineers and
others to gain a general scientific knowledge of artifacts similar in form
and substance to the scientific knowledge of nature developed by the
physical sciences. Metaphorically speaking, engineering knowledge was
freed from its jail of shape ."
The first recorded encounter between engineering and scale problems
arose after the invention of the torsion catapult in Classical Antiquity.
Starting from a well-constructed and successful catapult, a Hellenistic
engineer might very weIl have assumed that he could make an equally
good catapult with three times the power by simply building another
catapult three times the size of the first. For reasons that were not
obvious, attempts to scale-up catapults often failed; catapults built in
imitation of a model catapult but at a different scale often exhibited
unpredictable and undesirable changes in performance as compared with
the catapult used as a model. Since catapults were important new tools
of war, Hellenistic engineers sought to find a rule of design, such that
a catapult might be built at an arbitrary, predetermined size which would
demonstrate predictable performance. We are fortunate in having abrief
sketch of the early effort of Hellenistic engineers to und erstand and
control scale effects . It was written by a noted engineer, Philo of
Byzantium (c. 250 B.C .). According to Philo the HeIlenistic engineers:
. ..discovered that the diameter of the bore was the basic element, principle , and
measurement in the construction of artillery. But it was necessary to determine this diameter
not accidently or haphazardly, but by some definite method by which one could also
determine the proportionate measurement for all magnitudes on the instrument. But this
could not be done except by increasing or decreasing the diameter of the bore and testing the
results . And the ancients did not succeed in determining this magnitude by test, because their
trials were not conducted on the basis of many different types of performance, but merely
in connection with the required performance. But the engineers who came later , noting the
errors of their predecessors and the results of subsequent experiment s, reduced the principle
of construction to a single basic element, viz., the diameter of the circle that receives the
twisted skeins.'
Though the details are not completely c1ear, the ancient engineers
apparently used the method now caIled "parameter variation", an exper-
imental method long favored by engineers, though little used in the basic
sciences." This design effort, which amounted to a major engineering
development and research undertaking supported by HeIlenistic kings,
was successful in its immediate goal. 9
The HeIlenistic engineers understanding of dimensionality was limited,
however. They discovered no general relationships that would enable
engineers to predict the behavior of fuIl-sized prototypes by means of
experiments upon scale models. With the exception of the catapult, the
behavior of artifacts continued to change in unpredictable ways as they
were made larger or smaIler. The ultimate failure of engineers in the
ancient world to master the general problem of scale effects encountered
in experimentation with artifacts was expressed by Vitruvius, who wrote,
"In some machines the principles are of equal effect on a large and on a
smaIl scale; others cannot be judged of models." 10
These limitations persisted until weIl into the Nineteenth Century and
constituted a serious barrier to the development of a science or sciences of
engineering . Newtonian physics provided insights of enormous import-
ance, but the complexity of engineering reality defied even Newton's
reasoned simplifications.
NEWTONIAN PHYSICS ; ITS USE AND LIMITATIONS
Newtonian physics was a breakthrough in scientific thought which had

enormous implications for engineering. But it took some time for the
technological promise of Newton 's science to be fully realized in
engineering practice. Eighteenth-century engineers, such as CharIes
Hutton (1737-1823), used Newton as a basis for engineering mechanics,
involving the ability to construct idealized mathematical models of
various artifacts. Newtonian science was cast into analytical form and
developed by French scientists such as Jean le Rond D'Alembert (1717-
1783), and this powerful analytical mechanics had a large influence on
engineering in France and Europe generally. It would be hard to
overemphasize the value to technology of having a vocabulary, a set of
correct scientific concepts, and symbolic means of representing techno-
logical systems . 11
Idealization was important in the interactions of engineering with
Newtonian physics . At one level, the fact that mathematical models based
upon Newtonian science were idealized was an advantage; it made for
simplicity and ease of calculation for approximate results . But at another
level, the idealizations built into Newtonian physics constituted a barrier
to an exact analysis of engineering artifacts. Thus attempts to use
Newtonian physics to create a theory for engineering ran into problems
that were not fully understood until the 20th century. To get a correct
theory, engineers had to take into account shape and other dimensional
properties left out of Newtonian science . It is notorious that Newton was
less than fully successful in dealing with certain other topics, notably fluid
resistance, and his successors left fluid viscosity and other things out of
their theories, though the associated phenomena (such a turbulence) were
inescapable parts of engineering practice.
Clearly, a consistent mathematical model of physical reality.must
incorporate rules to ensure independence of scale, since otherwise the
mathematical model derived from physical theory might apply at one
scale, but not at all scales. Such a science would lack general validity.
Newton did in fact create a physics that was independent of scale . That
is, his theories were successful in predicting the behavior of a great range
of objects regardless of scale, as long a certain !imitations were
understood . Newton analyzed the scale problem and concluded that his
physics was valid at all seales in Proposition 32 , of Book Two of his
Principia, He considered the case where there are two systems of bodies
composed of an equal number of particles, where nothing is postulated
about their relative scale relationships. He then showed that if these two
systems are similar in geometrie relations, in motions , and in forces and
40 EDWIN T . LAYTON JR.
if they met certain other conditions (e.g. they are not in contact with one
another), then their behavior, and hence the laws of Newton's physies
would be independent of scale. This critieal proposition reads in part:
Suppose two similar systerns of bodies consisting of an equal number of particles , and let the
correspondent particles be similar and proportional, each in one system to each in the other,
and have a like situation among themselves , and the same given ratio of density to each
other; and let them begin to move among themselves in proportional times , and with like
motions (that is, those in one system among one another, and those in the other among one
another). and if the particles that are in the same system do not touch one another, .. .I say
that the particles of those systems will continue to move among themselves with like mot ions
and in proportional times."
It is important to realize, however, that the proof of this and other

propositions in Newton's Principia, rested upon a partieular ontology,
which Newton had developed earlier, notably in several propositions at
the end of Book One. In these he had shown that partieies acted as point
centers of mass with forces acting between these centers. He started with
the case of a sphere, and then extended his results to non-spherical bodies.
Though Newton's partieies had size (e.g. their diameters could be in
proportion to the forces acting on them) , they did not have shape; Newton
started with spherical particles , and showed that forces act at the point -
center . He then extended these results to include non-spherical bodies.
Thus even though a particle might be assumed to have a size, the shape
was irrelevant, since he had reduced material reality to point centers and
forces acting on these points. 13
Mathematieal points, of course, have no spatial or other dimensions .
In essence Newton achieved independence of scale by eliminating shape in
the fundamental entities of his system. The reality he described was
reducible to system of partieies and the forces acting between their point-
centers. Though geometrie relat ions among partieies were important,the
shape of the basic units was irrelevant, since they were effectively reduced
to dimensionless points . These propositions were part of a larger
ontology, whieh need not be discussed here, other than to emphasize the
fact that Newton achieved exactitude and scale independence (and some
other things) by this ontology . 14
Newton's ontology and with it his mechanisms for ensuring geometrie,
kinematie and dynamie similarity, broke down in the case of many
engineering problems. Indeed, Newton used his principle of similitude
(Proposition 32) to derive an incorrect theory of fluid resistance in
Propositions 33, 34, and 35. He attempted to reduce fluid resistance to

simple rules akin to his law of gravity. As modern fluid mechanics shows,
movement in a resisting medium is not that simple; one must take the
shape and other properties of the boundary layers into account. This
point is perhaps c1earest in Newton's Proposition 35, where he deals with
the drag of a sphere in a resisting medium, essentially the case from
which the Navier-Stokes equations were derived. The complexity of the
latter is in striking contrast to Newton 's theory. Newton avoided the
notorious, und still unsolved "n-body" problem, which arises in consider-
ing the mutual attractions of the many small particles for each other, by
using his principle of similarity he developed in Proposition 32 to show
that certain interactions of particles can be reduced to simple physical
principles. However, it should be noted that Newton, in Proposition 35
(and elsewhere in his Principia) is very clear about the assumptions
involved in his proofs."
Theodore von Karman showed that, with appropriate qualifications,
Newton's formula for fluid resistance could be applied to supersonic
flight , where the mutual attractions of the fluid particles can be
neglected." But no one condemns Newton for not having gone so far,
especially after accomplishing so much. His idealizations were historically
necessary and he provided solid foundations for future progress .
Charles de Borda (1733- I799) demonstrated the gap between
Newtonian theory and engineering experiment. Borda was a French
engineer and experimentalist interested in ship design and water wheels .
Both areas where Newton's theory was not adequate to the needs of
engineering. Borda tested Newton's theory of fluid resistance using a
rotating arm which pulled objects of various shapes through water; Borda
was able to measure the actual resistance. He found that Newton's
theoretical predictions were not verified by his experiments and he
conc1uded that the theory was so far from the truth that it would be
"useless and even dangerous to apply this theory to the art of constructing
ships . "17 Borda's experiments led to towing tank experiments on the
resistance to models of various forms of ship huIls, a line of study that, in
time, produced important information on dimensionality associated with
the names of Ferdinand Reech (1805-1880) and William Froude
(1810-1879) .18
John Smeaton (1724- I792) showed engineers how to use scale models
while avoiding the dimensional problems that long plagued such
endeavors. He played a pivotal role in the development of a methodology
42 EDWIN T. LA YTON JR .
for using scale models experimentally, particularly by his classic

experiments with models of water wheels and windmills. Smeaton's
success was based upon a clear understanding of the limits of model
experimentation. As he noted, "It is very necessary to distinguish the
circumstances in which the model differs from a machine in large;
otherwise a model is more apt to lead us from the truth than towards it.
Hence the common observation, that a thing may do very well in a model,
that will not answer in large. "19 Smeaton delayed publication of his model
experiments until he had compared their behavior with that of full-sized
artifacts.
Smeaton's work constituted the foundations of an engineering science
of turbomachinery based upon experiment. On the basis of his
experimental discoveries, Smeaton derived aseries of law-like relations,
or design principles, for water wheels and windmills. They had the form
of naturallaws, even though they were statements about man-made
devices. Smeaton was not clear about their ontological status. He called
them "maxims", that is, guides to prudent behavior, as in engineering
design. Smeaton's "maxims" were formulated in such a way as to directly
benefit the actual design of water wheels. In part because of this, his
research led to important innovations. In theory the power of all water
wheels depended on the weight of the water and the distance through
which it fell; there was no theoretical basis in Newtonian science for
thinking that the design of some water wheels or windmills would be
significantly more efficient than others. Smeaton, however, showed that
undershot wheels, in which water was applied at the bottom of the wheel
and acted by its impulse, were only about half as efficient as overshot
wheels, in which water was applied to the top of the water wheel and acted
by its gravity. As a result, overshot and breast water wheels replaced
undershot wheels in most British industrial establishments ."
Though Smeaton had opened the door, his procedures were severely
limited. His methodology rested on a sort of "survival of the fittest"
mechanism : only those relations that were independent of scale survived.
The others were discarded. The information discarded included the vast
majority of the valid scientific knowledge which could be gained by
experiment. Because of the limitations of Smeaton's methods, many
engineers insisted that experiments be conducted on full-scale artifacts .
For instance, Benjamin F. Isherwood, a leading American engineer of the
nineteenth century, held that "a fact, to be of practical authority in
engineering, must be derived from experiments made on the scale and
under the conditions of actual practice."21 This was , in fact, what was
done at the early textile center, LoweIl, Massachusetts, where James B.
Francis used only full scale turbines and other apparatus for his famous
hydraulic experiments conducted in the middle years of the 19th
century." Francis was in the forefront of engineers who were developing
experimental methodology and instrumentation which permitted an exact
comparison between theory and observation. Francis found that math-
ematical theories developed for engineering were not adequate guides to
engineering practice, a fact which he blamed upon idealization. For
example, Francis developed a weir formula to measure the flow of water
through reetangular openings . He was led to this successful investigation
by the inaccuracies of existing formulas . As he commented: "The result,
however, of these numerous labors, is far from satisfactory to the
practical engineer. On a careful review of all that has been done, he finds
that the rules given for his use , are founded on the single natural law
governing the velocity of fluids, known as the theorem of Torricelli;
omitting, in consequence of the extreme complexity of the subject, all
consideration of many other circumstances whieh it is weIl known,
materially affect the flow of water through orifices. "23
Francis found in the case of the turbine that mathematical theories
were no more accurate in dealing with the properties of these artifacts
than were experiments with scale models. The unreliability of theory led
Francis, to compile empirical tables giving, for turbines whose diameter
varied by one foot increments from 2 to 10 feet , the quantity of water
discharged, the horsepower and the rotary speed in revolutions per
minute, for turbines under heads varying from five to forty feet , at
increments of one foot. His data revealed serious errors in the available
theories, in rotor speed, in the rate of discharge, and in the power
developed. He commented that, "The turbine has been an object of deep
interest to many learned mathematicians, but up to this time, the results
of their investigations, so far as they have been published, have afforded
but !ittle aid to the Hydraulie engineer. "24
GRAPHICAL SOLUTIONS
Following Newton's work, it became evident that three kinds of similarity

needed to be taken into account to avoid scale dependence: geometrie,
kinematie, and dynamie sirnilitude."
The first, geometric similitude, was the first one to be fuIly achieved
through the descriptive geometry of Gaspard Monge (1746-1818), first
published in 1795. This was an important extension of Renaissance
projective methods and it was itself a fundamental engineering science
and also the starting point for many other important developments in
engineering science ." As it solved only one of the criteria for complete
similarity, scale effects remained and engineering was not freed from
dimensional constraints. Nevertheless geometric methods served
engineers amazingly weIl. Along with educated intuition and experience,
engineers managed to use graphics as a framework for profound scientific
advances in engineering. Graphics became a way ofthinking and knowing
as weIl as a means of communication, which has persisted until after the
middle of the Twentieth Century.
In the 19th century there was a strong tendency for the engineering
sciences to be expressed graphicaIly and mapped into descriptive
geometry, as, for example, in the cases of the graphical statics of Carl
Culmann, the circle diagrams of Otto Mohr, 27 the velocity triangles of
1ulius Weis bach and others for turbines, 28 and the indicator diagrams and
thermodynamic cycle diagrams for heat engines which began with the
work of lohn Southern and lames Watt. 29 Francis attempted to develop
a graphical method for the design of turbines; it was not completely
successful, but it revealed important facts that analytic theories failed to
reveal."
Indeed, engineers developed an entire sub -branch of graphics which
dealt with the graphical solution of equations which, in most cases , could
ha ve been solved equaIly weIl by standard mathematical methods.
American and also many British engineers came to feel comfortable with
geometry and to distrust analysis. lohn Perry, though a pupil of Lord
Kelvin, was an early mechanical engineering educator in Britain who
developed graphical methods for teaching the calculus to engineering
students." Methods of graphical solution of systems of equations are still
employed in engineering."
The success of graphical methods of analysis based on geometry was
probably due in part to the fact that geometry has similarity relations built
into its very fabric. We usuaIly think of the angles of triangles as measured
in degrees or radians. But if they are considered as the ratios of sides of
similar triangles, as in trigonometry, then they are also dimensionless
quantities. In general, a ratio of numbers with the same dimensions has
the effect of canceling out the dimensions. Since aIl mathematical
rnagnitudes can be expressed as ratios, pure mathematies is dimension

free. Dimensions arise only when mathematies is used to model physieal
or technologieal reality. Thus graphies carried more than just geometrie
similarity; mathematieal entities are inherently dimensionless . As such
they were guides to the discovery of other kinds of similarity, if these
could be expressed geometrieally .
Graphieal methods continued to be used in spite of the fact that they
too were idealizations. Engineers did not fully understand why their
graphieal methods sometimes worked so well, nor why they failed in other
cases. Engineers still relied on intuition in developing graphieal methods
appropriate to their discipline. But the graphieal staties developed by Carl
Culmann, for example, was a remarkable examples of how far graphieal
approaches could carry engineers, even in the absence of general methods
for avoiding dimensional traps ."
DIMENSIONAL ANAL YSIS
Though graphies provided hints, a full understanding of kinematic and

dynamie similarity required a new mathematieal theory of dimensions .
The explicit analysis of dimensions, appropriately called "dimensional
analysis", has played a large role in extricating engineering from its "jail
of shape" .34 The invention of the first dimensional analysis by Joseph
Fourier (1768-1830) was, therefore, an important event in the historieal
evolution of engineering science. The invention of dimensional analysis
was stimulated by the recognition that there was a need to ensure
dimensional homogeneity in physieal equations. Fourier's work appears
to be the first to deal with dimensional homogeneity explicitly. While
engineering was to be perhaps the principle beneficiary of dimensional
analysis, dimensional homogeneity was as necessary for physieal science
as for engineering.
Dimensional homogeneity is the situation when all the terms of an
equation have the same dimensions. As noted above, pure mathematies
had no problems with dimensional homogeneity, because there are no
dimensions. For applied theories, in whieh mathematies is used to model
physieal reality, the dimensions of real physieal quantities must be
assigned to mathematieal terms. Dimensions are measured in some kind
of units. Each symbol in an equation whieh models physieal reality will
stand for the product of a number with some units of measurement.
46 EDWIN T . LA YTON JR .
Though specific systems of units vary, as between the English and S.I. (or
International System) of units, for any given problem a set of
fundamental dimensions will be selected, such as mass, length, time, and
temperature, which are independent of the particular systems of units
employed. The number of fundamental dimensions is usually kept as
small as possible for greater simplicity. From the fundamental
dimensions, derived units can be constructed, such as force, energy, and
power.
To illustrate the principle of dimensional homogeneity, we can imagine
an equation as consisting of two or more terms that are to be added
together. Even though the entities which combine to form each product
may be different, the products to be added must have the same dimensions
raised to the same index. We can see velocity as the ratio of length and
time or [L] divided by [T]. For greater simplicity we avoid ratios and
express the dimensions as products raised to a certain power. Velocity is
therefore expressed the product [L] multiplied by [T-I ] . Since length,
time, and mass constitute an important set of fundamental dimensions,
they are often grouped together. In the above expression mass can be
expressed by capital M raised to the power zero.
We can express a hypothetical relation in which all of the products or
terms to be added have the same indices of the same fundamental
dimensions, for example, we find that the dimensions of all the terms can
be expressed by the dimensional product [MI] [U] [T-2 ] , where M is mass,
L is length, and T is time. These are, for example, the dimensions of
energy (and also of work and torque). If all the terms of an equation have
these same dimensions (after cancellation and simplification of the units
of dimension), the equation is dimensionally homogeneous.
If, however, we have an equation in which the terms to be added do not
have the same indices of the same dimensions, such an equation is said to
be inhomogeneous. We may suppose all but one of the terms in an
equation have the dimensions and their exponents just noted (for energy),
but that the term to be added to them has different indices of the
fundamental dimensions . For example, we can imagine a second term to
be added to the first in which the indices of dimensionality are different,
say those of velocity, [U] [MO] [T-I ] . In this case the equation is not
dimensionally homogeneous. Such an equation would be the symbolic
equivalent to adding kinetic energy to velocity, an operation having no
physical meaning.
Dimensional analysis was devised by Joseph Fourier as a byproduct of
his foundation of one of the first modern engineering theories, heat

transfer, in a c1assic work, Theorie analytique de la chaleur , published in
1822.3 5 Fourier was not an engineer but a scientist concerned with
applications. He considered his work as a further extension of c1assical
physics, more precisely, of the rational mechanics developed by Newton .
In a sense he was right in this, since he based his work upon the analytic
forms which Newton's ideas had received through French and other
mathematicians and scientists. At several places in his "Preliminary
Discourse" (or introduction) Fourier saw his work as a further extension
of rational mechanics and a necessary part of the science of nature. That
is, "the principles of the theory are derived, as those of rational
mechanics, from a small number of primary facts, the causes of which are
not considered by geometers, but which they admit as the results of
common observations confirmed by all experiments. "36 But one might
equally say that Fourier's work was very much unlike other branches of
Newton 's mechanics. Fourier emphasized the exceptional complexity of
heat transfer, in which he noted that his subject went beyond mechanical
theories. As he noted, " .. . whatever the range of mechanical theories,
they do not apply to the effects of heat . These make up a special order of
phenomena which cannot be explained by the principles of motion and
equilibrium. "37
In this introduction Fourier stressed the gains for pure science of his
new theory, though he was not unaware of its practical bearings. He
concluded his introduction with the words :
The theory of heat will always attract the attention of mathematicians, by the rigorous
exactness of its elements and the analytical difficulties peculiar to it, and above all by the
extent and usefulness of its applications; for all its consequences concern at the same time
general physics, the operations of the art s, domestic uses and civil economy .3.
Contrary to Fourier's expectations, his science of heat transfer became, in

the course of time, almost exclusively the concern of engineers and it is
now counted among the engineering sciences. It was thermodynamics, not
heat transfer, that attracted leaders in basic science. Fourier's invention
of dimensional analysis had a certain accidental quality. As Enzo
Macagno has pointed out, Fourier, in his early papers on heat transfer,
used the same two letters for four different terms, causing dimensional
inhomogeneity in his equations. However, it is noteworthy that when he
reworked his early papers into his c1assical book he further developed his
48 EDWIN T . LA YTON JR .
dimensional analysis while eliminating the problems that had occasioned

his earliest form of dimensional analysis. Dimensional homogeneity is a
general property of all equations and was not specifically tied to
engineering as a matter of logical necessity. But dimensional issues were
related to engineering, if only by practical contingency. Engineers found,
as did Fourier, that, theories dealing with the detailed, enormously
complex, detailed behavior of matter and energy required more variables
and this made is desirable to have an explicit test for dimensional
homogeneity ."
Fourier responded to the problem of dimensional inhomogeneity with
a stroke of genius: he invented a method of dimensional analysis to deal
with this problem. He gave it no name; it was included at the end of his
second chapter under the title, "General Remarks" . Fourier treated his
ingenious exponential algebra as a matter of little moment. It was, to be
sure, not comparable in mathematical sophistication or importance with
his invention of Fourier series in the same treatise ; but the method of
dimensional analysis has had a great importance for engineers and others .
The historical importance of a mathematical method is not necessarily
directly proportional to its mathematical depth and sophistication;
humbler discoveries may be of equal practical importance.
Fourier's invention was related to the fact that his theory was more
complex and required more variables than most other branches of applied
mechanics. Starting with four fundamental dimensions (length, time,
force [or mass], and temperature), he managed to reduce the number of
fundamental dimensions down to only three, by manipulating ratios of
the other dimensions. Fourier's reduction of his fundamental dimensions
to three had the effect of emphasizing the formal similarity of heat
transfer to classical mechanics, which used only length, time and force (or
mass).
However, heat transfer was a science that was not completely reducible
to mathematical abstractions such as lines, points, and angles. In
Fourier's case he had to use three more variables to characterize the
thermal properties of real materials. Fourier called these new terms
"specific elements" . In his own words:
The specific elements which in every body determine the measurable effects of heat are three
in number, namely the conductibility relative to the atmospheric air, and the capacity fOT
heat. The numbers which express these quantities are , like the specific gravity, so man y
natural characters proper to different substances."
HEAT TRANSFER AFTER FOURIER : INCREASING COMPLEXITY
Following Joseph Fourier a number of scientists and engineers con-

tributed to the development of dimensional analysis." It is significant that
some of the most important works using and advancing dimensional
analysis were those whose main thrust was to reduce or eliminate
unrealistic idealizations that had been incorporated into the fabric of
classical physics. Less idealization meant adding an increasing number of
terms and incorporating them in dimensionally homogeneous equations.
These efforts were remarkably successful; it turned out that these
mathematical abstractions, in spite of their idealized character, could
accurately reflect the phenomena observed in the real world .
As one of the founders of modern engineering heat transfer, Max
Jakob, pointed out many year later, one can classify sciences in terms of
the number of fundamental dimensions they require . Mathematics can be
conceptualized as having no dimensions or as having only one dimension,
length. Kinematics involves two dimensions, length and time, making it
intermediate between mathematics and mechanics . Mechanics requires
three fundamental dimensions, length, time, and force (or mass).
Thermodynamics adds temperature, but can leave out time if the goal is
to describe equilibrium conditions. Heat transfer, according to Jakob, is
inherently more complex than these other sciences, in that it requires four
fundamental dimensions, length, time, mass (or force), and tempera-
ture.?
Jakob may have overstated his case; the use of four or even more
fundamental dimensions is not unique to heat transfer. Nor does the
absence of dimensions from mathematics make the latter simpler than
kinematics. Hut he was correct in stressing the complexity of his
discipline. As he noted, since "these four dimensions can be combined in
three modes of heat transfer [conduction, convection, and radiation],
which are different in principle, it can be foreseen that heat transfer is a
rather complex section of science. "43
Fourier's theory was simple compared to the heat transfer theories that
were to follow. An important source of complexity in the engineering
sciences lies in the fact that these sciences serve not one but several fields
of engineering design. In response to the needs of design, as weIl as
internal scientific considerations, the number of applications and hence
of specific properties had grown enormously. For example an important
heat transfer text by Ernst R. G . Eckert and Robert M. Drake, published
50 EDWIN T. LAYTON JR .
in 1959, gave separate lists of properties and terms used for heat-
exchanger calculations, for thermal radiation, and for mass transfer.
Whereas Fourier had three "specific elements", heat capacity and two
measures of heat conductibility, later elaborations of heat transfer.
particularly convective heat transfer, had to take many more specific
elements onto account. In Jakob's Heat Transfer published in two
volumes in 1949 and 1955, there are nearly eleven pages devoted to
nomenclature. The 287 definitions of terms include some 59 which specify
specific properties of materials, energy fluxes, or characteristic lengths. In
addition Jakob listed 12 dimensionless parameters used in heat transfer."
In Eckert and Orake's Heat and Mass Transfer of 1959 there are five
pages devoted to nomenclature. Eckert and Orake list 55 specific
properties, along with fourteen dimensionless parameters needed for heat
transfer. The empirical richness of the field is further suggested by Eckert
and Orake's tables of property values given in an "Appendix of Property
Values".45
Enumerations of specific properties are inherently subjective. Indeed,
a certain design subjectivity is built into every engineering science. Some
specific properties may be divided into several related terms because of the
different machines or structures to which the theory is applied. In other
words, the number of "specific properties" depends on human goals and
the state of the art as weil as upon nature.
As engineering sciences have become more general and abstract, the
linkage with specific contexts was, inevitably, less weil defined . The
classification schemes adopted by contemporary engineering scientists in
the field of heat transfer seldom make the specific design context
completely clear. Eckert and Orake's reference to heat exchangers is
exceptional. Eckert became interested in heat transfer as an offshoot of
his interest in power plant combustion. In particular, he faced problems
of radiant heat transfer in the design of steam boilers fueled with
pulverized coal. In the index to his book Eckert makes no reference to
steam boilers or to coal. Eckert does, however, make brief references to
specific contexts in the text. The burning of pulverized coal in steam
boilers raised specific issues, some centered on "Beer's Law". The actual
design context, while not very explicit in the book, was very clear in
interviews of Eckert conducted by Layton and Goldstein.:"
As an analysis of the textbook by Eckert and Orake makes clear, the
proper context for understanding the engineering sciences is that of the
design of artifacts, though considerations internal to science are also of
DIMENSIONALIT Y AND ENGINEERING SCI ENCE 51
importance. Though the design context is often implicit rather than

explicit, it shows itself, in part, in the complexity introduced by the
number of specific properties which the theory must take into account if
it to be of value to designers of a wide variety of artifacts. Such sciences
are less idealized and more complex than theories which aim only at the
most compact and general explanations. In a basic physical science many
of the complexities of the engineering sciences can be avoided. The
proliferation of terms involves many considerations involving shape. For
example there are special symbols used for wall thickness, fin thickness,
entrance height. Of special interest are various characteristic lengths (such
as "shape factor", "equivalent radius", and "dimensionless wall dis-
tance") and a number of lengths associated with the boundary-Iayer
("boundary-Iayer thickness", "thermal boundary-Iayer thickness",
"flow boundary-Iayer thickness", and "boundary-Iayer displacement
thickness");? The proliferation of terms having reference to shape
indicate an ever-greater departure from a Newtonian ontology of forces
acting upon mathematical points. Such departures are likely to cause
dimensional complexities, and they account, in part, for the fact that the
engineering sciences often give a great deal of attention to dimensional
analysis.
A MATHEMATICAL TH EORY OF HYDROD YNAM ICS :

BACK TO SCALE -MODELS
The advance of dimensional analysis created new opportunities in other

sciences. This was particularly the case with hydrodynamics. As noted,
Newton's efforts to deal with movement in resisting media had not been
completely successful. 48 Viscous flow remained achallenge to physics and
engineering alike. A British scientist, George Gabriel Stokes (1819-1903),
sought to develop a correct mathematical theory of viscous fluid flow. In
this he built upon prior work by C.L.M.H. Navier (1785-1836), a
distinguished French engineer and pioneer in engineering science. Stokes
did not change the form of Navier's equations, but he introduced
important new concepts in the new way in which he derived what came to
be called the "Navier-Stokes" theorem. To eliminate unrealistic ideal-
ization, Stokes introduced a key dimensional idea, that of "dynamic
similitude", that is similarity in forces, in his derivation of the "Navier-
Stokes" equation."
Dynamic similitude was a master key to understanding similarity

relations. It led to striking advances by a number of figures, including
Hermann von Heimholtz (1821-1894) , lohn William Strutt, Lord
Rayleigh (1842-1919), and Osborne Reynolds (1842-1912). Heimholtz
became one of the founders of modern model theory in a rem ar kable
paper published in 1873. He examined the ratios of various forces relevant
to the problem, anticipating the use of many similarity conditions that
were to be of subsequent importance. In examining the ratios of force he
addressed a basic problem of the modern model engineer in considering
how to model the wind resistance of dirigible balloons. The problem came
from the fact that a smaller model does not mean smaller wind velocities.
Heimholtz discovered a relation between the size of the model and the
velocity of the wind. The wind velocity for a small model is now called the
"corresponding speed". Helmoltz showed that, for dynamic similarity,
the velocity varied inversely with the linear dimensions of the model. Since
the model was assumed to be much smaller than areal balloon, the
velocity of the air had to be much greater than that encountered in actual
practice by the full-scale artifact. This meant that for small models the
velocities for air must be very high, possibly exceeding the speed of sound.
Heimholtz, showed how this situation could be modelIed by letting the
model move in: water rather than air. In this case the greater density of the
water allows dynamic similarity to achieved at lower velocities."
Rayleigh was among the first physicists to use Fourier's dimensional
analysis. He developed a purely algebraic method to normalize his
differential equations in order to make them homogeneous in
dimension." This method of ensuring the dimensional homogeneity of
equations is entirely sound, and it is still widely used, particularly by
physicists. It had the enormous advantage of linking the analysis of
dimensions to the physical realities of the problem, as expressed in the
governing equations. In Rayleigh's hands it transcended the issue of
dimensional homogeneity; it proved to be a powerful tool for solving all
manner of problems, even deriving relations (such as the sag-span ratio
employed in the design of suspension bridges) that others had laboriously
determined by experiment. Rayleigh's approach, powerful and funda-
mental as it was, had an unanticipated disadvantage. It caused Rayleigh
to miss the nature and importance of dimensionless parameters. In fact
Rayleigh had used the inverse of the Reynolds number without realizing
its significance, since he used it just as an arbitrary mathematical
expression needed to make his equations homogeneous in dimension. 52
Though he respected Reynolds, Rayleigh treated Reynolds' principal

work in dimensionality as (mainly) an application of Stokes's concept of
dynamic similarity (which Rayleigh had further developed)."
Osborne Reynolds' work was of immense importance for the
development of the conception of dimensionless parameters and their use
in engineering theory. He published alandmark paper in 1883. It was
indeed possible to derive the dimensionless parameter later termed the
Reynolds number from considerations of dynamic similarity. But this
deduction did not prove anything about the flow . 54 Reynolds set out to
understand turbulent flow. He reasoned that , "they [the governing
equations of flow] might contain evidence which had been overlooked, of
the dependence of the character of motion on a relation between the
dimensionalproperties and the external circumstances of motion." He
then derived a dimensionless parameter and commented that , "this is a
definite relation of the exact kind for which I was in search.... It
seemed ...to be certain, if eddies were due to one particular cause, that
integration would show the birth of eddies to depend upon some definite
value [of this parameter]. "55
Reynolds still could not immediately convince hirnself that turbulent
flow was a consequence of viscosity, but he finally concluded, in a
remarkable paper on turbulent flow in 1895, that: "when an indefinitely
smalI ...disturbance exists, it will fade away if [the Reynolds nu mb er] is
less than a certain number. .. .While for greater values of this
function ...the energy of sinuous motion may increase.. ." .56 That is, he
demonstrated that this number, the rat io of inertial to viscous forces,
really did mark the transition from laminar to turbulent flow. 57 The
elucidation of the role of dimensionless parameters was of crucial
importance for engineering theory, and in this task Reynolds was a very
important pioneer. 58
Reynolds made an important contribution to constructing model laws.
Reynolds devised "distorted" scale models of hydraulic systems such as
the Mersey River, which exaggerated certain phenomena to make them
more apparent. 59 Reynolds made many other important contributions,
for example in contributions to the theory of lubrication, where he built
on experimental data accumulated by another engineer, Beauchamp
Tower."? He also contributed to heat and mass transfer in a paper in which
he made an important analogy between the turbulent exchange of
momentum and heat." But it is above all Reynolds' discovery of
dimensionless parameters and his conceptualization of them as ratios of
54 EDWIN T . LA YTON JR.
forces that laid the foundations for future engineering work in

dimensional analysis.
TOWARDS A UNIFICATION OF ENGINEERING SCIENCES
In the first quarter of the Twentieth Century a few pioneers saw possibil-
ities for the unification and systematization of engineering sciences by
means of dimensionless parameters. The engineering theory of the
turbine, for example, remained unsatisfactory because certain key terms
could not be calculated accurately by the application of mechanics,
notably the rotor speed and the amount of water used. The first of these,
the "specificspeed", was discovered by Rudolph Camerer in 1905. 62 He
discovered the capacity coefficient and other dimensional fundamentals
which enabled him, with a number of other insights, to unify and
systematize the theory ofturbines by 1914. 6 3 On further investigation, the
specific speed turned out to be determined by the shape and configuration
of certain key components of the turbine. For example, In the Pelton
impulse turbine the specific speed is determined by the ratio of the
diameter of the rotor to the diameter of the jet of water that propels it; in
a Francis reaction turbine the geometry of the rotor vanes is a critical
determinant of specific speed. In this context it is clear why a purely
mechanical theory failed to make accurate predictions of rotor speed."
Wilhelm Nusselt was one of the most important of the unifiers of heat
transfer theory. Nusselt had earned his doctorate in mechanical
engineering at the Technische Hochschule in Munich under Oskar
Knoblauch . Knoblauch trained three of the foremost founders of
modern, engineering heat transfer: Nusselt, Ernst Schmidt, and Max
Jakob. Schmidt was in turn the teacher of Ernst R .G .Eckert. 6 5
The study of convective heat transfer was one of the most important
tasks undertaken by these pioneers in heat transfer. After Ludwig
Prandtl's founding of modern fluid mechanics and boundary layer
theory, it became possible to write the equations of convective heat
transfer, but they could not usually be solved. As John Lienhard has
shown, Nusselt was a pivotal figure in the development of the theory of
convective heat transfer. He was the first to apply dimensionless
parameters outside hydrodynamics, and in doing so he helped to
illuminate the correlations among groups of parameters and consequently
their larger role in the formulation of engineering theories . In aseries of
papers culminating in 1915, Nusselt showed how the functional form of

the solutions could be derived and the limited experimental data
correlated by means of dimensional analysis . In particular Nusselt
showed how convective heat transfer could be simplified and coordinated
by basing its theory upon aseries of dimensionless parameters which were
functionally related to each other. Besides the Reynolds number, Nusselt
derived the dimensionless parameters latter to be named the Nusselt,
Prandtl, and Grashof numbers, and showed their functional interdepen-
dence."
What Nusselt had done for one part of one engineering science, Edgar
Buckingham, a physicist with the United States Bureau of Standards, had
already proposed for all engineering sciences in a fundamental paper
published in 1914 in which he set forth his "7C' theorem.":" Buckingham's
7C' theorem is sometimes treated simply as a way of simplifying physical
theories by reducing the number of fundamental terms . In this part of his
work Buckingham was anticipated by Amie Vaschy, a French telegraph
engineer. As a result the theorem is sometimes called the Vaschy-
Buckingham theorem." Both Buckingham and Vaschy saw that for a
system of R variables in N dimensions, it was usually possible to describe
the system completely by functions of R-N dimensionless products. More
formaIly , the simplification made possible by the Vaschy-Buckingham
Theorem, is now phrased as follows : the number of products in a
complete set of dimensionless products of the variables X I , X2, • • • ,X n , is
n-r, where r is the rank of the dimensional matrix of the variables. To
Vaschy this was essentially all there was to his theorem; it is this
simplification with which he begins his paper, and their is no discussion of
the general theoretical implications of his theorem; the rest of his
excessively short paper is a list of examples of the application of his
theorern."
The statements of Buckingham's 7C' theorem which emphasize only the
simplification, neglect the most fundamental of his points from the point
of view of the history of the scientific development of engineering. We can
think of the Buckingham 7C' theorem as consisting of two parts. In his most
fundamental paper in 1914 he deals successively (in numbered sections)
with the most general form of physical equations (Section 1) in which he
concludes that in the most complete form, the coefficients of the terms of
a physical equation are dimensionless numbers. In his second section
Buckingham introduced Fourier's conditions of dimensional homogene-
ity, and then introduced and defined his dimensionless 11" terms and
showed that there were fewer in number by the number of fundamental

dimensions (equal to n-k terms). Buckingham then went on to show how
a physical equation might be converted into a function of 7l" dimensionless
parameters. In his third section he gave exampies of how to form the
dimensionless 7l" terms for specific cases." Buckingham's statement is
much fuller than Vaschy's, and his emphasis upon the theoretical
meaning as weIl as the practical consequences of his theorem is clear and
explicit. Though Buckingham's 7l" theorem is an important improvement
over Vaschy's, he was careful to claim no credit for originality, and may,
in fact, have been guided by a reading ofVaschy's paper which he put into
more complete, clear, and rigorous form."
Vaschy's paper appears to have had little influence. Buckingham's
papers had an enormous impact in America, though apparently less in
Europe where Rayleigh's methods predominated . Buckingham clearly did
see the importance of his theorem , and he strove, with considerable
success, to point out its utility to engineers.? The Vaschy-Buckingham
theorem stimulated a great deal of engineering research on various aspects
of dimensionality. Model experimentation became a rigorous discipline in
the hands of a number of researchers. Henry L. Langhaar noted that
Buckingham's theorem revealed that, in the ideal case, where "complete
similarity " can be achieved, the 7l" terms of the model and prototype are
equal to each other. In practice this is rare ; model and prototype seldom
behave in exactly the same fashion . Consequently, as Langhaar observed,
" an important part of the work of the model engineer - indeed the most
important part - is to justify his departures from complete similarity or to
apply theoretical corrections to compensate for them ."73 But Langhaar
was among the foremost in showing that Buckingham's theorem helped to
unify engineering, notably by showing that dimensional analysis is
intimately related to model experimentation and model theory .
One of the main lines of modern advance beyond Buckingham has
been in relating the dimensionless parameters to the physical conditions of
the problem (e.g. to the governing equations). This can be seen as
synthesizing the method of Rayleigh (which started with the governing
equations of the problem) with the methods of Buckingham and Vaschy.
Stephen J. Kline's emphasis upon certain fundamental ratios of forces,
and other aspects of his analysis had been enormously enlightening ."
Kline expanded upon Reynolds' idea of the dimensionless parameter as a
ratio . Kline saw that dimensionless parameters were not just arbitrary
constructs, but they (or at least some of them) were fundamental ratios
that expressed important truths about the world. It turns out also that
these are the sort of truths that are of great value for engineers attempting
to understand the detailed behavior of matter. H.A. Becker has called
Kline's approach the "Pythagorean method" and also, synonymously,
"the method of ratlos" . By whatever name, Kline's approach is
fundamental in that it combines dimensional analysis with the physieal
analysis of the problem to provide a richer and more powerful tool for
engineers. A notable feature of Kline's work, as extended by Becker is the
emphasis upon physieally significant ratios of forces, dimensions, fluxes,
and the like. Kline's work is a remarkable synthesis of much of the
concepts and procedures related to dimensional analysis, put into a more
meaningful physical context. 75
The result of the research by Kline and others building upon
dimensional analysis has been a much deeper understanding of the role of
dimensionality in engineering. In some early works the 7r terms were
referred to as "variables" . Kline has stressed the important point that the
terms whieh are combined to constitute dimensionless ratios are no longer
variables but parameters. This is one of the more significant results of the
investigations of dimensionality . These parameters reveal a fixed set of
dimensional relationships, expressed in the 7r terrns, which define the
dimensional structure of the problem (and indeed of a system of problems
sharing the same dimensional structure). Once this underlying
dimensional structure of the problem or system is properly expressed, it
does not vary; when it does vary one is going from one class of problems
with a partieular dimensional structure to another dass of problems with
a different dimensional structure."
H.A. Becker is an outstanding representative of those engineers and
others who have sought a deeper understanding of dimensionality by
means of the mathematieal theory of groups. This line of analysis led
Becker to examine related families of dimensionless parameters and to
classify them by the nature of the physieal entities whose ratios they
express: force, flux ratios, charge ratios, and geometrie ratios . Following
Kline, Becker arranged related clusters of these ratios on grids whieh
reveal many important physieal relationships and interconnections, in a
manner reminiscent of the periodie table of chemieal elements. It is
significant to note that this methodology places configuration in the
foreground. Becker wrote that, "the idea is akin to that of a system, but
with emphasis placed on the 'shape' of the system and the events that
occur within it. "77
Though the reference is not to the shape of the physical objects of

engineering science, the emphasis upon configuration and "shape" in the
structure of the theory is certainly suggestive. Peter Kroes and A.
Sarlemijn have pointed out (from the perspective of the philosophy of
science and technology) the important role of analogy in the technical (or
"engineering") sciences. Using the mathematical theory of groups to
analyze analogies in science and technology, they note that there is a
difference in the types of analogies used in physics and in technical (or
"engineering") sciences." Kroes has emphasized the importance of
analogies between physical systems, which he calls "structural analogies" .
These analogies include scale models and mathematical models of
engineering systems. It is in the very nature of engineering artifacts that
they constitute systems of components and subcomponents, and physical
systems models are needed to represent them." As Kroes has concluded:
" ... especially in the technological sciences, structural analogies between
systems are often more important than structural analogies between laws.
Within the technological sciences there is a whole body of knowledge
known as 'theory of similitude' .. . which is focussed on the analysis of
structural analogies between technologically relevant systems. "80 This
very promising line of inquiry thus further suggests that shape or structure
is a fundamental element within engineering sciences, both at the level of
artifacts and also at the meta-level of theories about theories of artifacts.
The maturing of dimensional analysis meant that engineering
experiment and engineering theory could be brought into close agree-
ment. The rapid growth of the engineering sciences required the
development of special instrumentation of all sorts. This topic is too
broad for treatment here. For heat transfer alone an impressive array of
new or improved measuring instruments were developed by engineers and
others for use in developing systematical experimental traditions in heat
transfer." As an example of new or improved instruments, the Mach-
Zehnder interferometer is particularly striking and important. It has
evolved, mainly since the 1930's, as aprecision tool used both in high
speed aeronautics and in heat transfer. 82
The result of the rapid advance in engineering research in the
Twentieth Century has been a convergence in the engineering sciences.
This convergence rests upon kinship in the basic formal structure of the
theories developed in fluid mechanics, heat, mass, and momentum
transfer, thermodynamics, and many special theories, such as those of
combustion, turbine blade cooling, ablation cooling of the nose cones of
rockets, fluid film lubrication, aerodynamics and many others.? The

rapid advance of engineering theory owed much to the development of
appropriate dimensionless parameters; by 1959 there were 210 of these in
use." They allowed large areas of engineering to be placed into the same
sort of mathematical form as other physical theory. This allowed a
powerful body of mathematical and experimental tools to be applied to
engineering problems, the long sought goal of scientific engineering.
The dimensional advances in engineering have produced a convergence
between physical science and engineering as weIl as convergence among
engineering disciplines. There is an increasing flow of people, ideas, and
instrumentation between basic and engineering science. Engineering
theory and experiment have become closer to physical science; this is
precisely what one should expect by the restructuring of engineering as a
"mirror image twin" of the basic sciences.
ONE SCIENCE . TWO CUL TU RES?
Percy Bridgman, a philosophically-minded physicist, developed the

theory of dimensional analysis . Dimensional analysis sometimes gives the
illusion of giving something for nothing and leading to deep, meta-
physical insights. Some authors thought to find the truly "fundamental"
dimensions through dimensional analysis , but Bridgman attacked the
illusion of "fundamentality". He showed that there are no fundamental
dimensions: all dimensions are defined by operations which are ultimately
arbitrary. Convenience and social needs determine what dimensions are
taken to be fundamental. Dimensions are the products of culture, and
reflect human needs and values; they are not written into the nature of the
universe .85
There was a famous exchange between Joseph Fourier, one of the
principal founders of the modern engineering science of heat transfer, and
Carl G. J. Jacobi, one ofthe foremost exponents of "pure" science. It was
appropriate that Fourier was the founder of heat transfer; though not an
engineer, he was an ardent exponent of the idea that science should serve
the practical benefit of humanity. This led hirn to make an intolerant
attack on Jacobi's devotion to pure mathematics, particularly for
allegedly "wasting" time in the study of elliptic equations. Jacobi's reply
to Fourier's criticisms is justly famous. "It is true," Jacobi maintained,
"that Monsieur Fourier had the opinion that the principle aim of
mathematics was public utility and the explanation of natural

phenomena; but a philosopher like hirn should have known that the sole
end of science is the honor of the human mind, and that under this title a
quest ion about numbers is worth as much as a question about the system
of the world. "86
Modern engineering c1early honors the human mind just as the basic
sciences have joined in serving the goals of public utility, We are liberating
technology from its jail of shape. In an age of renewed environmental
concerns and energy shortage we will need all the scientific power we can
muster to meet modern challenges . Whether we have learned to use this
enhanced power with greater wisdom than in the past remains to be
determined.
University 0/ Minnesota
NOTES
I V. Sackville-West, 'The Land - Summer' , in Collected Poems (London : Webb and

Bower, 1989), p. 69. Reprinted courtesy of Curtis Brown and John Farquharson. I would
like to thank Professor Rutherford Arris for calling my attention to th is poem and for his
inspiration and help over many years .
, On the diameter of tubes in relation to their mechanical properties, see Charles
MacGregor, ' Mecha nical Properties of Mate rials' , in Theodore Baume ister et al. (eds.) ,
Marks ' Standard Handbook for Mechanical Engineers (New York: McGraw-HiIl, 1978),
Sec. 5, pp. 49-50. For a cylindrical tube the circumferential or "hoop" stress, S, is equal to
S = pr/t where pis the internal pressure, r the radius and t the plate thickness .
J On "specific speed" see W.G. Whippen, ' Hydra ulic Turbines' , in Baume ister , Marks '
Standard Handbook for Mechanical Engineers, Sec. 9, pp . 137-138 . The discover y of

specific speed will be discussed below . One of the most impo rtant keys to rat ional design of
water wheels was to avoid "angle of incidence" losses due to shock at entrance. To eliminate
shock it is necessary to have "tangent entr y" of the water, which , in turn , required that the
designer knew the speed of the rotor.
4 Robert L. Daugherty and Joseph B. Franzini, Fluid Mechanics with Engineering
Applications (New York: McGraw-HiIl, Seventh Edition, 1977), pp . 445-446.
, J .J . Cornish, 'Aeronautics' , in Baumeister, Marks ' Standard Handbook for Mechanical
Engineers, Sec. 11, pp. 60-64, 66-71. In the most obv ious case, in laminar flow the
boundary layer flows as if made up of many thin layers moving smoothly parallel to one
another; in a turbulent boundary layer the flow is no Ionger parallel; there are irregular
motions (such aseddies) normal to the surface of the layer .
6 There is a modern tendency to restriet the meaning of "dimensional analysis" to the
contents of the Vaschy-Buckingharn 'Ir theorem, but a much bro ader defin ition is used here.
I ha ve been much influenced by Stephen J . Kline, Similitude and A pproximation Theory
(New York : McGraw-HiIl, 1965), who uses "dimensional analysis" in a broad sense. He
analyzes four distinct types of dimensional analysis (pp. 262-215). These are (I) the 11"
theorem (discussed below), (2) the use of governing equations, that is, normalizing the
governing differential equations to make them dimensionally homogeneous. This method
was pioneered by Lord Rayleigh; it is favored by physicists, and (3) the "method of
similitude" , (which is largely Kline's own creation). To this one should add the method of
"configurational analysis" which is built upon Kline's work, but which emphasizes the use
of group theory for the analysis of dimensional issues. (For configurational analysis, see
H .A. Becker, Dimensionless Parameters (New York : John Wiley, 1976), Vol. 3, pp. 10-11.
To these I add historical methods of analyzing dimensions such as that of Fourier, who
invented the first formalism of dimensional analysis. In addition, I follow Kline in including
the derivation of modellaws within dimensional analysis (pp . 2-7). Modellaws are often
called mies or principles of similitude (or similarity). Other works dealing with dimensional
analysis that I found particularly helpful include Percy W. Bridgman, Dimensional Analysis
(Carnbridge, Mass: Harvard University Press, 1921); W.J . Duncan, Physical Similarity and
Dimensional Analysis (London: Edward Arnold, 1953); H.E. Huntley, Dimensional
Analysis (London: McDonald, 1953); Henry L. Langhaar, Dimensional Analysis and
Theory 0/ Models (New York: John Wiley, 1951); and L.1. Sedov, Similarity and
Dimensional Methods in Mechanics (New York : Academic Press, 1959). On modelling
theory I found Langhaar especially useful. Several of these contain useful bibliographies,
notably Becker and Kline. Kline has comments on rnajor contributions to the literature
which can be found in the text . On the history of dimensional analysis I am deeply in the
debt of Enzo O. Macagno, 'Historico-Critical Review of Dimensional Analysis', J. FrankIin
lnst. 232(1971), pp. 393-394. Macagno 's study is a critical and informative history of
dimensional analysis; Macagno follows the more common practice of restricting the
meaning of "dimensional analysis" to the Vaschy-Buckingham theorem and closely related
matters.
7 Philo of Byzantium quoted in Morris R. Cohen and I.E. Drabkin, A Sourcebook in Greek
Science (Cambridge, Mass.: Harvard University Press, 1958), p. 318.

• On parameter variation see Walter G. Vincenti, "The Air-Propeller Tests of W. F. Durand
and E. P . Lesley: A Case Study in Technology Methodology', Technology and Culture 20
(October, 1979), p. 714.
• The traditional phrase "Research and Development" ref1ects the now outdated "applied
science" theory of technology, which has now been replaced by the "interactive" model.
There is evidence suggesting that it is more common for research to follow the appearance
of a novel engineering (or design) idea, For example, one of the largest and best of the
innovations studies was done by the well-known economists Sumner Myers and D. G.
Marquis, who found that successful innovations in the industrie s studied did not originate
with research in 95 percent of the cases, but came about in implementing a technological
idea. (See Sumner Myers and D.G. Marquis, Successful lndustrial Innovations (Washington
D. C.: The National Science Foundation, NSF 69-17, 1969)especially Table 19, p. 46.) On
the interactive model see Barry Barnes, 'The Science- Technology Relationship: A Model
and a Query ', Social Studies 0/ Science 12 (February, 1982), pp . 166-171 and Edwin T .
Layton, Jr ., 'Mirror Image Twins: The Communities of Science and Technology in 19th-
Century America' , Technology and Culture 12 (October, 1971), pp. 562-580 . Studies of
technology as knowledge include Edwin T. Layton, Jr ., "Technology as Knowledge' ,
Technology and Culture 15 (January, 1974), p. 41; Waller G. Vincenti, What Engineers
Know and How they Know It, Analytical Studies from Aeronautical History (Baitimore:
Johns Hopkins, 1990).
' 0 Marcu s Vitru vius Polio, The Architecture 0/ Marcus Vitruvius Polio in Ten Books,
transl. Joseph Gwilt (London: Priestley and Weale, 1826), p. 342.

11 CharIes Hutton, A Course 0/ Mathema tics, 2 vols . (New York : Samuel Campbell and
others, American edition edited by Robert Adrian based on 5th and 6th London editions,
1818). Volume 11 is an applied mechanic s for engineers which used a simplified Newtonian
notation. "Hutton" seems to have been very influential in the English speaking countries .
It went through a total of29 editions, of which nine were American, between 1798 and 1860.
(National Union Catalog 0/ pre-J956 Imprints 262, pp. 184-186 .) Though still used in
English speaking countries as late as 1860, Hutton's work was made obsolete by the French
tradition which employed Leibnitz 's notation and shifted mechanics from a geometric to an
analytical formalism . For a broad survey of mechanics which stresses the European and
French perspectives, see Rene Degas, Histoire de la Mecanique (Neuchatel: Editions du
Griffon, 1950).
12 Isaac Newton, Mathematical Principles 0/ Natural Philosophy and His System 0/ the
World, 2 vols., transl. Florian Cajori (Berkeley: University of California Pre ss, 1962),
Vol.Il , pp . 327-328 . See also Proposition 33, Theorem 26 including the several corollaries
(pp . 328-331) (hereafter: Newton, Principia).
IJ Newton, Principia, Book I, section 12 deal s with spherical bodies , section 13 extends the
analy sis to non-spherical bod ies, pp . 193-225. See particularly Proposition 88, Theorem 45,
" If the attractive forces of the equal particles of an y body be as the distance of the places
from the particles, the force of the whole bod y will tend to its center of gra vity; and will be
the same with the force of a globe, consist ing of similar and equal matter, and having its
center in the center of gravity." (I, p . 216).
'4 Newton's ontology also included his "geometrization of space" , and othe r th ings. These
issues have long been debated. Though now somewhat outdated, historically important
work s in understanding Newton 's ontology (or metaphysics) included Edwin A. Burtt, The
Metaphysical Foundations 0/ Modern Science: A Historical and Critical Essay (New York :
Harcourt Brace, 1925) and Alexandre Koyre, Etudes Galileennes (Paris: Hermann, 1939)
and the same author's , Metaphys ics and Measurements: Essays in Scientific Revolution
(London: Clapham and Hall, 1968).
.. Newton, Principia, I, pp. 328-336. Newton's awareness of the limits, and possibilities, of
his ontology is remarkable and evident in these propositions of fluid resistance . Newton did
deal with some of the effects of the mutual interactions of the particles of a fluid, in
Proposition 33, pp . 328-329 , where he shows that the centripetal and centri fugal forces by
which the particles of the system act on each other, can be reduced to simple considerations
of the for ce and matter as with a simple system by using the similarity principles developed in
Proposition 32. Thus Newton refuted Descartes , while not solving the problems of Fluid
Mechanics with which modern engineers must deal. The fact that Newton's physics was
associated with a particular ontology does not mean that it was invalid except with that
ontology. Thus it is possible (with later analytical tools) to expand Newton 's F = MA to get
the Navier-Stokes equation, and by considering the ratios of two such expression s for two
different systems to derive the key dimens ional concepts of dynamic and kinematic similarity.
Thi s is done, for example, by Max Jakob, Heat Trans/er, 2 vols. (New York : John Wiley,
1949, 1955), Vol. I, pp . 16-19,429-430. It would be a-historical to expect Newton to have

done this; the fact that it can be done is a tribute to the foundat ions which Newton provided
for physical science. The Navier-Stokes theorem will be discussed below.
" Theodore von Karman, ' Isaac Newton and Aerodynamics', J. Aeronautical Sei. 9
(December, 1942), pp . 521-522, 548.
17 CharIes de Borda , ' Experiences sur la resistance des fluides' , Memoires de l'Academie
Royale des Seiences (1767), pp . 495-503. Borda's tests were repeated and extended using a
towing tank (so that the model did not move through disturbed water) and the towing tank
became a significant advance in model testing methodology. Some other conflicts between
theory and experiment in the 18th century are summarized in Terry S. Reynolds, Stronger
Than a Hundred Men. A History 0/ the Vertical Water Wheel (Baltimore : Johns Hopkins,
1983), pp. 204-223, 237-241.
" Froude's relationship to Ferdinand Reech and the story of the law of similarity and the
dimensionless parameter named after Froude is complicated and will not be discussed in this
paper . As Rouse and Ince note, Froude did not originate the similarity rule (it was discovered
by Reech), and Froude never used the "Froude Number" . On the other hand he did anticipate
boundary layer analysis. See Rouse and Ince, History 0/ Hydraulics, pp. 154-155, 182-187.
I. John Smeaton, 'An Experimentallnquiry Concerning the Natural Powers of Water and
Wind to Turn MiIIs, and Other Machines, Depending upon Circular Motion ' , Phil. Trans.
Royal Soc. 51(1, 1759-1760), pp . 100-101. Smeaton's role in the experimental tradition in
engineering is discussed in Terry S. Reynolds, Stronger than a Hundred Men, pp. 218-233.
20 Reynolds, Stronger than a Hundred Men , pp. 223-226, 280-284.
" Quoted in Edwin T . Layton, Jr ., 'American Ideologies of Science and Engineering',

Technology and Culture 17 (October, 1976), p. 693.
" Edwin T . Layton, Jr ., 'Scientific Technology: the Hydrau lic Turbine and the Origins of
American Industrial Research', Technology and Cu/ture 20 (January, 1979), pp. 72-78, 81.
Jame s B. Francis , Lowell Hydraulic Experiments (New York: D. Van Nostrand , second
edition , 1868).
IJ Francis, Lowell Hydraulic Experiments, p. 52; the experimental derivation of the Francis
weir formula is on pp. 69-135 .

24 Francis, Lowell Hydraulic Experim ents , p. 52 and Table 4, pp . 53-54. The properties
which Francis measured turned out to be dependent upon dimension , as will be noted below
in discussing specific speed.
25 One can infer from Newton's Proposition 32 basic geometric, time, and force similarity
conditions, where land 2 stand for similar systems, and S, T and F are any lengths, time
intervals, and forces occurring in the similar systerns, the relation s S,/S2 = a, T ,/T2 = b , and
T ,/T2 = c, where a, b, and c are factors of proportionality. A common alternative for this
system of units is to consider length, time and force instead of mass.
2. Peter Jeffrey Booker, A History 0/ Engineering Drawing (London : Chatto & Windus,
1963), pp. 18-36,86-106.

27 Stephen Timoshenko, History 0/ the Strength 0/ Materials (New York : McGraw-HiII,
1953), pp. 190-197,283-288. Both of these are systems of analyzing the stresses within a
framework by graphical means. Culmann, a mathematician by training , became convinced
tha t structural calculat ions could be best done graphically. See Carl Culmann, Die
graphische Statik (Zurich : Meyer und Zellers Verlag, 1869).
2. Julius Weisbach, A Manual 0/ the Mechanics 0/ Engineering (New York: John Wiley,
Trans. A. Jay Du Bois, 1880), II, pp. xliii-xlv, 222-226, 321-326, 355-358, 397-400.
2. D.S.L. Cardwell, From Watt to Clausius (Ithaca : Cornell University Press, 1971), pp.
80-83 .
30 Francis, Lowell Hydraulic Experiments , pp. 36-43 .
31 Larry Owens, ' Vannevar Bush and the Differential Analyzer: Text and Context of an
Early Computer' , Technology and Culture 27 (January, 1986), pp. 88-89. John Perry,
Calculusfor Engineers (London, 1897).
32 Philip Franklin, 'Oraphical Representation of Functions' , in Baumeister, Marks'
Standard Handbook for Mechanical Englneers, Sec. 2, pp. 54-62 .

33 There is an interesting discussion of graphical statics and the link between this scienceand
design in David Billington, Robert Maillart's Bridges. The Art of Engineering (Princeton :
Princeton University Press, 1979), pp. 5-7
34 As noted above, 1 am much indebted to Kline and follow his very broad definition of
dimensional analysis. He categorizes dimensional analysis as apart of "fractional analysis"

(e.g., analysis which provides useful but not complete information) and sees the latter as a
fundamental methodology in engineering. See Kline , Similitude and Approximation
Theory, pp. v-vii, 2-7, 87-219 .
J5 1 have used Alexander Freeman 's English translation: Joseph Fourier, The Analytical
Theory of Heat (New York: Dover reprint of the 1878 English edition, 1955).
36 lbid. , p. 6.
37 lbid. , 2. This is true for heat transfer where rates (that is time) are critical; but the later
science of thermodynamics can be considered in terms of equilibrium, without concern for

the rates at which the transfer of heat takes place.
38 lb id., pp. 23-26 .
39 Macagno, ' Historico-Critical Review of Dimensional Analysis', pp. 393-394. The

further developrnent of heat transfer added a good deal of complexity to Fourier 's theory .
It would be surpris ing if the first investigator in a new science discovered everything of
relevance and importance. Fourier's theory was primarily useful in conduction; though he
made beginnings in convective and radiative heat transfer, these latter theories were less
satisfactory . Modern heat transfer uses the concept of thermal diffusivity in the study of
heat conduction, and especially the Fourier number, a dimensionless parameter needed to
correlat e information in unsteady heat conduction . (E.R.O. Eckert and Robert M. Drake,
Heat and Mass Transfer (New York: McOraw-HiII, 1959), p. 77).
40 Fourier, Theory, p. 127. See also pp. 126-130. The variables Fourier admitted were x
(length), t (time), v (temperature), c (heat capacity) , and two heat transfer coefficients (in
Fourier's nomenclature) , "specific conductibility", K, and "surface conductibility", h (p.
130). See also Jakob, Heat Transfer, Vol. I, pp. 3-5 .
41 Macagno, 'Historico-Critical Review of Dimensional Analysis', pp. 391-402. See also
H.E. Huntley, Dimensional Analysis (London : McDonald, 1952), contains abrief history
(pp. 33-44) . For another historical study see Alton C. Chick, "The Principle of Similitude',
in John R. Freeman (ed.), Hydraulic Laboratory Practice (New York: the American Society
of Mechanical Engineers, 1929), pp. 796-797.
42 Max Jakob, Heat Transfer, Vol. I, pp. 3-5.
43 lb id., I, p. 5.
44 lbid., I, pp. 716-726.
4. Eckert and Drake, Heat and Mass Transfer, pp. x-xiv, The append ices of properties are
located on pp. 493-521.

•• Edwin T . Layton, Jr. and Richard J . Goldstein, 'Interviews with Ernst R. G . Eckert" ,
(five videotaped interviews, Walter Library, University of Minnesota, Minneapolis,
Minnesota) .
• 7 Eckert and Drake, Heat and Mass Transfer, pp. x-xiv.
•• Newton, Principia, Book Il. See also Rouse and Ince, History of Hydraulics, pp . 83-86.
•• George Gabriel Stokes, 'On the Effect of the Internal Friction of Fluids on the Motion of
Pendulums', in George Gabriel Stokes, Mathematicaf and Physicaf Papers, 5 vols .
(Cambridge: Cambridge University Press, 1880-1905), Vol. III , pp. 1-141, especially pp. 7-
17. The broader historical context of the Navier-Stokes equation is sketched by Rouse and
Ince , History of Hydraulics, pp. 193-200. Dominique Noir has argued that the Stokes law
of similitude for the drag of a sphere in a resisting fluid can be deri ved from Aristotle's
physics . See Dominique Noir , ' La prerniere loi de similitude de 1a rnecanique des fluides ',
Revue de l'Histoire des Seiences et de feurs Applications 25 (1972), pp. 271-274.
50 Hermann von HeImholtz, ' Über ein Theorem, geometrisch ähnliche Bewegungen
flüssiger Körper betreffend , nebst Anwendung auf das Problem, Luftballons zu lenken',
Monatsber. Kön. Preuss. Akad. Wiss. Berlin 188 (1873), pp . 501-514. This is a rem ar kable
for eshadowing of dimensional analysis in which he anticipated the later similarity
considerations asso ciated with the Froude, Mach, and Reynolds criteria; there is no
evidence, however, that he anticipated the existence or the role of dimensionless parameters.
For the more modern modelIing ofthe Helmholtz's problem see Edgar Buckingham, ' Model
Experiments and the Forms of Empirical Equations', Trans. Am. Soc. Mech. Engineers 37
(1915), pp. 273-277.
51 Lord Rayleigh (John William Strutt), ' T he Principle of Sirnilitude', Nature 95 (1915), pp .
66-68, 644 is a very remarkable virtuoso performance. Kline has a remarkably clear and
insightful chapter in which he analyzes the cornparatives ad vantages and drawbacks of
Rayleigh's method with that of Buckingham and others. (See Kline, Similitude and
Approximation Theory, pp . 262-215 .) For a succinct statement, see also Murdock.
'Mechanics of Fluids' in Baumeister, Marks' Standard Handbook for Mechanicaf
Engineers, Sec. 3, p. 50.
52 Rayleigh 's use of dimensional analysis and his use of the inverse Reynolds number is
discussed in Macagno, ' H istorico-Critical Review ofDirnensional Analysis' , pp . 395-396.

53 Lord Rayleigh (lohn WilIiam Strutt), 'On the Que stion of the Stability of Flu ids' ,
Scientific Papers, 6 vols. (New York: Dover Reprint of 1869-1919, 1964), Vol. IV , p. 575.
5' Jakob, Heat Transfer, pp . 429-431 . Reynolds is not cornpletely clear on his reliance upon
Stokes in this paper .

55 Osborne Reynolds , ' An Experimental Invest igat ion of the Circumstances Which Deter-
mine Whether the Motion of Water Shall Be Direct or Sinuous and the Law of Resistance
in Parallel Channels', Papers on Mechanicaf and Physicaf Subjects , 3 vols . (Cambridge:
Cambridge Univ. Press, 1900-1903), Vol. Il, pp . 54-55 . See also Rouse and Ince, History
of Hydraulics, pp . 206-211, which puts Reynolds' work in perspective.
5. Reynolds , 'On the Dynamical Theory of Incompressible Viscous Fluids and the
Determination of the Criterion', Papers, Il, p. 543.

57 lbid., pp . 51-105 .
5. Macagno, ' H istorico-Critical Review of Dimensional Analysis ', pp . 395-396. Macagno

points out that Fourier used a dimensionless term, and that he was aware of its dimensionless
character (p. 394). Though Reynolds was very irnportant, the concept of the dimensionless
parameter and its role in engineering was the work of many investigators, and only gradually
crystallized in the period 1890 to 1920.
•• Reynolds, ' On Certain Laws Relating to the Regime of Rivers and Estuaries and on the
Possibility of Experiments on a Small Scale' , in Papers, H, pp. 326-335. See also his three
committee reports on scale-modeling in hydraulics, ibid., pp. 351-380.
60 Reynolds, 'On the Theory of Lubricat ion and its Application to Mr. Beauchamp Tower's
Experiments ...' , ibid., pp. 228-310.

6 ' Reynolds, 'On the Dynamic Theory of Incompressible Fluids and the Determination of
the Criterion ' , and ' On the Extent and Action of the Heating Surface of Steam Boiler', ibid. ,
pp. 81-85 . This analogy linked heat with mass transfer; it is discussed in Ernst Eckert , 'Heat
Transfer' , in D.M. McDowell and J .D. Jackson (eds.), Osborne Reynolds and Engineering
Science Today (Manchester : Manchester University Press, 1970), pp. 160-175. McDowell
and Jackson's anthology gives a good idea of the scope of Reynold's work.
62 Rudolph Camerer , 'Klassificat ion von Turbinen', Z. Ver. Deutsch. Ing. 49 (4 March,
1905), p. 308.
63 Rudolph Camerer, Vorlesungen über Wasserkraftmaschinen (Leipzig: Wilhe1m
Englemann , 1914). Camerer was assisted by the fact that American turbine companies
produced stock turbines at standard sizesand these were tested at the Holyoke Testing flume
at constant head, so that entire families of geometrically similar turbine could be compared
in their speed and their "swallowing ability" (capacity) at constant head (p. 295). For the
capacity coefficient of turbines, see Denis G. Shepherd, Principles ofTurbomachinery (New
York: Macmillan, 1956), pp. 31-32.
64 Daugherty and Franzini , Fluid Mechanics with Engineering Applications, pp. 445-446,
468-469,490-493 .
.. John H. Lienhard, 'Ernst Kraft Wilhelm Nusselt', Dictionary of Scientiflc Biography,
s.v. See also John H . Lienhard , 'Notes on the Origins and Evolution of the Subject of Heat
Transfer', Mechanical Engineering 105 (June, 1983), pp. 20-27 . See also Elizabeth Jacob,
'Max Jakob; July 20, 1879-January 4, 1955', and Virginia Dawson, 'Frorn Braunschweig to
Ohio: Ernst Eckert and Government Heat Transfer Research', and Ernst R.G. Eckert,
'Ernst Schmidt - As 1 Remember Hirn', in Edwin T. Layton and John H. Lienhard, eds.
History of Heat Transfer (New York: American Society of Mechanical Engineers, 1988),
pp. 87-116 , 125-137, 139-146.
66 John H. Lienhard, ' Nusseit' , and ' Notes on the Origins and Evolution of Heat Transfer'.
Wilhelm Nusselt, 'Das Grundgesetz des Wärmeüberganges' , Gesundheits-Ingenieur 38

(October, 1915), pp. 477-482. Nusselt's correlation of the Nusselt, Prandtl, and Grashof
was reviewed by Jakob, Heat Transfer, I, pp. 481-488. On the overall impact of Nusselt's
similarity methods and the Nusselt number on heat transfer, and Nusselt's other publi-
cations of importance see (in addition to Lienhard), Jakob, Heat Transfer" I, pp. 496-499,
522,527-528,744. Ernst R. G. Eckert thought Nusselt's paper so fundamental that he wrote
that its publication in 1915marked "the birth year of a scienceof heat transfer.' (Eckert and
Drake, Heat and Mass Transfer, p, 229).
67 Edgar Buckingham, 'On Physically Similar Systems: Illustrations of the Use of
Dimensional Equations', Phys. Rev 4 (1914), pp. 345-376.

68 Aime Vaschy, 'Sur les lois de similitude en physique', Ann. Telegraphiques 19 (1892),
pp. 25-28 and the same author's application of his principle in 'Sur les lois de similitude de
electricite', ibid., pp. 189-211. Macagno has useful remarks on Vaschy for which I arn,
again, in his debt; he points out that Vaschy did not make it completely explicit that the
terms he dealt with were dimensionless, though he did note that his terms were independent
of the system of units employed (a characteristic of dimension1ess numbers).
•• Aime Vaschy, 'Sur les lois de similitude en ph ysique", p. 345. Vaschy's statement is very
compressed; his statement of his theorem takes less than a page, and he moves on to
illustrations and applications at the bottom of the first page . At the outset of the theorem
he states tha t • "The most generallaw of similitude in mechanics and physics results in the
following theorern." He then states that a function of n physical quantities (a" ai ... •an ) can
be reduced to another of (n-p) parameters of the form fix" X2 •.., x n - p ) in which the
parameters x" X2 x n_p are monomial ("single term") functions of a" a2•.... Vaschy does not
state that this is the most general or fundamental form for physical equations, only that it
is possible . Vaschy , in his first statement, makes no statement suggesting that the x
parameters are dimensionless. Later on in the paper, however, he does say that these terms
ar e independent of the units in which these terms are expressed . This is today adefinition of
a dimensionless parameter, but perhaps in 1892 these useful entities were not so weil known
and an explicit statement might have helped some readers. In any case Vaschy and his fellow
French telegraph engineers do not appear to have made wide use of the theorem . One
problem appears to have been Vaschy's extreme brevity .
7. Buckingham, 'On Ph ysically Similar Systems ' , J. Wash. Acad. Sei. 4, pp . 345-350 (Sec.
2. pp . 345-346). As Langhaar has shown , the mere fact of dimensional hornogeneity leads
directly to the 11" theorem. (Langhaar, Dimensional Analysis, pp . 55-58). In some early
works the 11" terms were called variables , but the term parameter is more correct. The
dimensionless parameters are composed of terms which , in the non-dimensional form , are
correctly called variables. But the process of dimensional analysis does more than ju st group
together "variables" into groups in which the units of dimension cancel out.
11 In his paper 'Model Experiments and the Forms of Emp irical Equations', p. 265,
Buckingham presented a short version of his theorem which is strikingly similar to Vaschy's
earlier statement of the theorem.
" Buckingham's paper 'Model Experiments and the Forms of Empirical Equations", pp .
273-296, presented to the American Society of Mechanical Engineers in 1915 was a
particularly striking and influential example of Buckingham's concern to popularize his
work among engineering. See also Buckingham, 'Physically Similar Systems ' , pp . 347-353
and his 'T he Principle of Similitude', Nature 96 (December 9. 1915), pp . 396-397, and his
' Notes on the Dimensional theory of Wind Tunnel Experiments', Smithsonian
Miscellaneous Collections 62 (4, 1914). pp. 15-26 .
73 Langhaar, Dimensional Analysis and Theory 01 Models , p. 64. Buckingham discussed
model experimentat ion in his ' On Physically Similar Systems', pp . 369-372. and with many
practical examples in his ASME paper, ' Model Experiments and the Forms of Empirical
Equations' , pp . 273-296.
,. Kline, Similitude and Approximation Theory.
,. For remarkable examples of the power of Kline's methods see his Chapter 4, pp . 68-199.
7. Kline, Similitude and Approximation Theory, pp. 74-75 .
77 Becker, Dimensionless Parameters , p . 6.
" A. Sarlemijn and Pet er Kroes, 'Technological Analogies and Their Logical Nature'. in
P .T . Dublin (ed.), Philosophy and Technology 4: Technology and Contemporary Life
(Dordrecht, Holland: Reidel, 1988), pp . 237-255.

7' Peter Kroes, 'Structural Analogies between Physical Systems ', Brit. J. Phil. Sei. 40
(1989), pp . 145-154. Structural analogies include both physical models and dimensional
analysis applied to governing equations. As Kroes points out , these structural analogies
between physical systerns (or the equations describing them) are not the same as the nomic
isomorphism of laws .
80 Ibid., p. 153.
8\ Ernst R.G . Eckert and R.J. Goldstein (eds .), Measurement Techniques in Heat Trans/er
(London: The Advisory Group for Aerospace Research and Development, NATO, 1970).
This book does not include the Mach-Zehnder Interferometer, which , however, is discussed
by the editors in numerous publications .
81 On the Mach-Zehnder interferometer see W . Kinder, 'Theorie des Mach-Zehnder
Interferometer und Beschreibung eines Gerätes mit Einspielgeleinstellung", Optik I (1946),

pp . 413-448. On the role of Ernst R. G. Eckert in this development, see Virginia P . Dawson,
'From Braunschweig to Ohio: Ernst Eckert and Government Heat Transfer Research', in
Layton and Lienhard, History 0/ Heat Trans/er, pp . 125-134.
83 Günter Küppers, 'On the Relation between Technology and Science - Goals of
Knowledge and Dynamics of Theories. The Examples of Combustion Technology,
Thermodynamics and Fluid Mechanics', in Wolfgang Krohn, Edwin T. Layton, Jr. , and
Peter Weingart (eds .), The Dynam ics 0/ Seience and Technology (Dordrecht : D. Reidel ,
1978). Sociology of the Seiences Yearbook, Vol, 2, pp . 113-133 .
8< E.F. Boucher and G.E . Alves , ' Dimensionless Nurnbers', Chem , Eng. Prog. 55
(September, 1959), p . 55. The result of dimensional hornogeneity, as shown much later, was
that engineering theories were then expressed as functions of dimensionless terms.
8> Bridgrnan, Dimensional Analysis. Bridgman's proof of the 1f theorem is given on pp. 36-
47 . Bridgman set forth his " operatio nalism" and his refutation of the idea of absolute
"fundamental" dirnensions, pp . 10-25.
86 Quoted in EricTemple Bell, Men o/Mathematics(New York : Simon and Schuster, 1937),
p.338.
PETER KROES
ON THE ROLE OF DESIGN

IN ENGINEERING THEORIES;
PAMBOUR'S THEORY OF THE STEAM ENGINE
INTRODUCTION : DESIGN AND THE NATURE OF

TECHNOLOGICAL KNOWLEDGE
This paper deals with the role of design in engineering theories. I

Staudenmaier considers engineering theories to be one ofthe four charac-
teristics which distinguish technological from scientific knowledge.! He
gives the following definition of this type of theory, which he claims is
more less generally accepted:?
An engineering theory is a body of knowledge using experimental methods to construct a

formal and mathematically structured intellectual system. The system explains the
behavioral characteristics of a particular dass of artifact or artifact-related materials .
Engineering theories are considered to be different "in both style and

substance" from scientific theories." According to Layton and Vincenti
the special cognitive status of engineering theories is due to the fact that
technological theories deal with man-made devices, or artifacts, whereas
scientific theories concern "natural objects" . The intention of this paper
is to investigate how this specific feature of the intended domain of
application of a theory, namely the fact that it deals with artifacts, affects
the structure of a theory.
It must be noted from the start that the opposition between natural
objects and artifacts does not really discriminate science from technology
as regards their object of study. Almost all phenomena studied by modern
experimental science are artificially created in laboratories and are in that
sense man-made, just as the objects of technology are. Thus, technology
as weIl as science both study artifacts, because they share the experimental
approach. There is, however, an important difference in the kind of
artifacts and the perspective from which they are analyzed. The artifacts
studied in technology have a (direct or indirect) technological function
and are analyzed from that point of view; that is, this technological
function has to be described and explained in terms ofthe artifact's design
and construction , From a scientific point of view, the possible techno-
69
P. Kroes and M. Bakker (eds.), Technological Development and Science in the Industr ial
Age ,69-98.
70 PETER KROES
logical function of an artifact is, however, irrelevant and not taken into
account; artifacts studied in science are not necessarily technologically
relevant, or are isolated from their technological context. Apparently,
this circumstance may lead to the development of different kinds of
theories within a technological and scientific setting .
THEORIE
OE LA
MACHINE A VAPE UR.
lh,"'l illt' prou"ll'r I'in u l clilud. d" metbode. e n uu r pout e.aluu 10:'
...sfc l. lIU le. proportio n, ..I" muhi"" ~ ".' ,", ur ; 0:1 j, Y , u b' til ut"·
11IIt leri. d o (ortft ul.. C1111"yli'1ue• • pto p"'•• de"'rm int1" I. " i l tu~
·r un e machin e de enee ,ou t IIno cbu co ee ne ue • •• charc c p.or u ne
l'i1eue üeee , " ' 3 potisuion ro ll t d ~. eIfeIl 'DUltl' I U. fcree en ehe .
'-3101 '. \011 t!Tel ut tle raUf 1111" ~o n ,o m m ' li o n een nue d'n u Cl de com -
IIuu ilJI1', 1.1 c.hmr r,e 0101 I ~ de len h' 'lu"il ("'1,1 1 lui do n ne r PO" t l ui (:li,.·
I'rooln i...· I nn m.H imum cI'elTe' uu le , ete • • ('tc. ;
n:n' l) 'c: ~
AP I' END ICE

..
L .. . . ~. I I A. ~ l" ._ "' . _ • •• ~ ... ,.... &._10., ~ In .
• lu L.."1....• . ..t .~ • •~ 1• • 1 , ... .... r u l. i•• • • • , .u., .,'.....1. '· "., . f ' .
" An L E C DEV. F .-;\1. G. DE PAiUDOUR l

..... Cl l'\ u.nl ltI L' U:O L& POLTT IC.~IQ tl ' ,
" .. 11" " 1 1 1-::11 1\ u , f' lIl1lr ' '' 'I YAt":l tl l. L' ,uni LLu l l l , "r; l "t "I · ""' '':
PARIS ,
IlACHELJEH, Il\IPRll\lEUH-LIBRAIRE
rou. LU U ....T lliKAT JQUU .
QU. \1 DE$ .\U G UST I I'lS , 55.
1839
Title page of the first French edition of Pambour's book .

PAMBOUR'S THEORY OF THE STEAM ENGINE 71
THEORIE
DE LA
MACHINE A VAPEUR.
gO o .
CHAPITRE PRDlIER.
PREUYES 01:: L'I:iEXACTITUOE DES ~JtTllOOES
ORDlNAIRES OE CALCUL.
§ .". Mode de calcul en usa{;c [usqu'lci,

Le but de cet ouvrage est de dernontrer que le
calcul des effets ou des proportions des machiues
io vapeur. tel qu'iI est ernploye dans la prariqu- ,
Oll indique par les au teurs qui on t trai te ce suje r •
est entierernenr Iautif, et de deve lopp er une nUII-
velle rheorie qui mene io des resul tats exacrs, ~ous
devons donc nec essairernent co nsacrer le premier
chapitre a prouver l'inexactirude des methodes
ordinaires dc caleu l , et nous passerons ensu ite
successivement au developpernent de la theor ie
a
proposee et son app lication aux divers systernes
.Ir machirres a vapcllr en usage.
L'effer produit par une machine se co rnpose
de deux elements : la resistallce q u'el le met CIl
mouvement " t la vitesse qu 'ellc communique d
cet te resistance. 11 en resulte que les calcul s qui
I
Beginning of the first chapter of Pambour' s book .
Pambour's theory of the steam engine has been selected as an example

of an engineering theory. This theory is a mixture of scientific and
technological principles. It fits the above definition perfectly weIl.
CardweIl describes Pambour's theory as:"
a paradigm of exact, critical engineering study : an exhaust ive investigation of the generat ion
of steam and its progress from the boiler through the cylinder to the condenser or
atmosphere. All types of engines were considered and the treatment was satisfactorily
analy tical.
Whether or not Pambour's theory is representative of engineering theories

72 PETER KROES
in general is a matter which will not be considered here. In order to portray

the technological nature of Pambour's theory it will be compared to
Carnot's theory of the steam engine. A comparison of the aim of these
theories shows a conspicuous difference between the two , Pambour sets
out to solve certain (design) problems concerning a particular dass of
artifacts, namely steam engines characterized by a specific design. Carnot,
on the other hand, is interested in developing a general theory about the
transformation of heat into work in any conceivable heat engine; the
concrete design of the artifact (heat engine) does not play any role. The
argument developed in this paper is that the prominent role of design in
Pambour's theory confers a distinctive technological flavour on this
theory. It will be shown that the design involved plays a crucial role in the
demarcation 0/ the domain 0/ application 0/ the theory, in one 0/ the two
main principles 0/ the theory, and in the kind 0/ concepts employed.
In the following, the notion of design will be used in the sense of a
technological design, that is a pattern or scheme showing how a particular
technological function or aim may be realised (in principle) . The fulfilment
of a technological aim or function under specified input conditions and
boundary conditions is the rationale underlying any technological design;
this means that the purpose of every element in a design and their mutual
arrangement is derived from the overall technological function. The input
and boundary conditions together with the technological aim define in
general the specijications of a technological device or artifact. For a given
design these have to be related to relevant design parameters; design
parameters are the "internal degrees of freedom" in a particular design .
The values of design parameters have to be fixed such that the required
specifications are met. Different designs with different design parameters
rnay, of course, fulfil the same technological function .
Descriptions ofthe function, design, specifications and construction of
an artifact may make use of concepts which derive their meaning primarily
from the technological context of the artifact. Such concepts will be
referred to here as technological concepts. Examples of technological
concepts in the case of the steam engine are: foot-pounds per minute,
bushels of coal, boiler, cylinder, cylinder pressure, piston, crank,
condenser, horsepower etc. These technological concepts do not
necessarily have any scientific (physical) significance. The following
analysis of Pambour's theory illustrates that the possibility of translating
technological concepts into scientific ones is of primary importance for an
understanding of the interaction between science and technology from an
PAMBOUR 'S THEOR Y OF THE STEAM ENGINE 73
epistemological point of view. Pambour is able to employ scientific

theories in solving technological problems on the basis of an embedding of
certain technological concepts (such as "load of a steam engine",
"cylinder pressure") into a scientific discourse.
PAMBOUR'S THEORY OF THE STEAM ENGINE
Francois Marie Guyonneau, le Comte de Pambour (1795-?), graduate of

the Ecole Polytechnique, published his famous book Theorie de la
a
machine vapeur in 1837. A few years earlier (1835), he had published a
successful book on steam locomotives. The book Theorie de la machine a
vapeur became a standard work in the field of steam engine theory for
many decades, went through several editions and was translated into
English and German." "As late as 1876", Kerker remarks, "it was author-
itatively referred to as 'the most celebrated treatise of de Pambour. ..
published in 1844, then far superior to other works and still in many
respects one of the best standards on the subject'''. 7 Pambour's work was
discussed extensively in the same year by Clausius in Die mechanische
Wärmetheorie." By that time it had become clear , on the basis of
thermodynamical considerations, that some of the assumptions made by
Pambour were not valid . The most important shortcoming of Pambour's
theory was his assumption that steam remains saturated during expansion.
In the thermodynamic studies of the steam engine by Rankine and Clausius
it was assumed thar condensation of steam takes place during expansion.
In view of the importance attached to Pambour's work by his
contemporaries, it is rather rem ar kable that his work has received little
attention in the history of steam technology . He is, for instance, not
mentioned in Singer's A History of Technology" nor in Dickinson's A
Short History of the Steam Engine'" and he is mentioned only once in
Matschoss's Die Entwicklung der Dampjmachine etc. 11 and Hills's Power
jrom Steam ;" Cardwell devotes less than one page to Pambour's work in
From Watt to Clausius." The only significant secondary sources on
Pambour, known to the author, are from Kerker . 14 It seems that the low
profile of Pambour's work in the history of steam technology is due to the
fact that it is not relevant to two dominant approaches in this field, namely
the "nuts and bolts" and the "science oriented" kind of history of
technology.
As will become clear below , Pambour's theory is not a thermodynamic
74 PETER KROES
theory of steam engines. It seems likely that Pambour knew of Carnot's

work, but he does not refer to it. 15 Against the background of the
development of thermodynamics, Pambour's book may indeed be
characterized as an "orthodox treatise" . 16
Pambour's book was intended for engineers as weIl as foremen
("Werkmeister") (p, 7).17 It offers a thoroughly mathematical treatment
of the steam engine. It starts with a critical discussion of the inadequacies
of existing theories and exposes a new "general theory" of the steam
engine . Pambour's theory makes it possible to caIculate the (net) work of
a steam engine under different operating conditions, as weIlas its velocity
and the required evaporation capacity ofthe boiler. This general theory is
then applied to and elaborated for various types of steam engines,
incIuding atmospheric engines and single or double acting high pressure
engines with or without condensation and expansive action. For each type,
Pambour derives a number of practical formulas for computing, for
instance, the velocity of the piston, the net effective power and the amount
of water evaporated per minute. For several types he also determines the
conditions for maximum efficiency .
Apparently, the determination in advance of the power of steam
engines could stilI not be solved in a satisfactory way in the time of
Pambour. For Newcomen engines the design tables drawn up during the
18th century had been an invaluable tool, but they were of no use for the
new kinds of steam engines (especially the high pressure engines)
introduced at the beginning ofthe 19th century. 18 In the introduction to his
book Pambour writes (p. 7):19
Although one did believe that it was possible to compute in advance the performance of an
engine that was to be built, experience showed that one could be confident about the required
power only in those cases where the engine was a fair copy of an already existing machine, and
that an y deviation made the power to be expected uncertain.
According to Pambour, the existing theory for computing the net power of
steam engines, which he refers to as the Ccefficientenrechnung (p. 9) was
defective. Roughly, it proceeded in the foIlowing way. It assumed that the
pressure in the boiler was the same as the pressure in the cylinder. From this
pressure, the back pressure (or the atmospheric pressure) had to be
subtracted, resuIting in the effective pressure P . Then, given the velocity v
of the piston, the "theoretical" power W of the engine was computed; W
= Pav, with a the surface of the cylinder . In practice, the power delivered
always turned out to be much less. It was standard procedure to estimate

the actual power W' by introducing for each type of engine a fixed
coefficient k the value of which was established on the basis of experience;
then W' was equal to kW. For instance, for stationary high pressure
engines without condensation k was assumed to be 0.50 (p. 14); but k could
be as low as 0.25 (p. 16). The difference between the theoretical and the
actual power of an engine was considered to be due to several circum-
stances, among which were friction, loss of steam and condensation of
steam (p , 15).
Pambour rejects the coefficient approach for a number of reasons .20 In
his opinion, it is based upon wrong theoretical presuppositions, the most
important of which is that the pressure in the boiler is the same as in the
cylinder, or, more generally , that there is a fixed proportion between the
pressure in the cylinder and in the boiler. He also describes experiments
from which he infers that the value of k cannot be constant but is different
for the same engine, or for engines 01 the same type working under
different conditions. Moreover, this theory leads to absurd estimations of
the amount of power lost by friction; in some cases more than 50070 of the
total power produced would have to be consumed by friction, whereas
measurements indicated that that figure should be much less (about 10 to
20%) (pp. 14-16).
Apart from these defects, it was not possible, on the basis of the existing
theory, to determine the velocity of an engine, given the resistance it had to
overcome. Pambour claims that his theory does not suffer from these
shortcomings. In particular, it offers a solution for the following problems
(p . 8) :21
Two points are of importance for an engine : the resistanceit puts into motion, and the ve/ocity
with which this is done. First of all, therefore, the following two problems turn up .
First . When the velocity is determined for an already existing engine : to find the resistance
that the engine can move forward with th is velocity.
Second. When the resistance is determined for an already existing engine : to find the
ve/ocity with which the engine will put th is resistance into motion.
Here the expression "for an already existing engine" implies that the
evaporating capacity of the boiler is supposed to be known . These two
problems are directly related to the determination of the power of steam
engines, for the product of resistance and velocity is equal to the power
("(Nutz)-wirkung"). Pambour's theory also addresses another kind of
problem [ibidem) :22
76 PETER KROES
But there is still a third problem emerging from the two previous ones , namely :
Third, When the resistance and the velocity are determined : to find the dimensions of the
engine that will be able to overcome this resistance with the given velocity . In the case of steam
engines, what is being searched for here is the size ofthe boiler, or, ifyou like, the evaporation
capacity corresponding to the required power.
This is clearly a design problem. As formulated by Pambour, it concerns

tbe determination of tbe capacity of the boiler given the desired power of
an engine." Note tbat for Pambour tbe most important design criterion
with respect to the desired power output is apparently not the size of the
cylinder (the area ofthe piston), as was tbe case for Newcomen engines, but
the evaporating capacity of the boiler. The reason for this becomes clear
wben we bave a closer look at bis tbeory.
Pambour's tbeory is founded on two basic assumptions (p. 25). Tbe
first assumption is that, as soon as an engine bas reacbed astate of constant
motion, dynamic equilibrium requires that tbe motive force must be equal
to tbe resistance or total load. Otherwise, the motion would be accelerated
or decelerated;" Tbis condition immediately implies that in general the
pressure of tbe steam in the cylinder is not equal to or a fixed proportion
of tbe pressure in the boiler, but is determined by tbe resistance (p. 22):25
Consequently, the moment the steam from the boiler enters the cylinder, its pressure changes
and becomes equal to the resistance of the piston . This circumstance alone constitutes the
foundation of the theo ry of steam engines and explains all that is necessary.
Therefore, any tbeory, sucb as tbe Coefficientenrechnung; which assumes

that tbere is a invariable ratio between tbe pressure in the boiler and in the
cylinder is necessarily wrong . Tbe second assumption is tbat oll the steam
generated in the boiler bas to proceed tbrougb tbe cylinder . In otber words,
tbere is no loss of steam, not tbrougb the safety valve, not through
condensation, or otberwise. In the following, the first assumption will be
referred to as the "principle of dynamical equilibrium", the second as tbe
"principle of tbe conservation of steam" .
Pambour constructs his theory on tbese two cornerstones. First he
considers tbe most simple case, that is an engine in which the temperature
of tbe steam is constant, and which pro duces rotative motion without
expansive action oftbe steam (pp. 23-25). In that case, it is a simple matter
to solve tbe tbree problems stated above. If PI is the pressure in tbe
cylinder and R the resistance per unit of surface on tbe piston (including
friction), then the first principle implies:
PI = R (1)
The second principle allows the computation of the velocity v of the piston.
Let S be the amount of water evaporated per unit of time in the boiler, and
q the number by which S has to be multiplied in order to get the correspond-
ing volume of steam at pressure P in the boiler. Assuming that the
temperature ofthe steam does not change, Boyle's law can be applied to get
the volume of steam of pressure PI (in the cylinder) generated per unit of
time ; this volume is equal to
(2)
This must be equal to the volume swept through by the piston per unit of
time; that gives
qSP
av =-- (3)
PI
with a the surface of the piston. Using (1) we arrive at:
qSP
av =-- (4)
R
For this simple case, this equation solves the three problems Pambour
started out with . If the velocity is known, the resistance can be computed
or vice versa (assuming that S is known), and if the desired power is known
(vR) then S follows immediately!
The derivation of a similar equation starting from more realistic
assumptions turns out to be much more difficult. The most general
situation considered by Pambour is a steam engine with condensation
working expansively (p. 175). In such an engine, the temperature of the
steam is not constant. The pressure and the temperature change when the
steam leaves the boiler and enters the cylinder, and also during the
expansion phase. Thus, Boyle's law cannot be used to calculate the change
in pressure of the steam entering the cylinder or during expansion.
Nevertheless the relation between pressure and volume of the steam in the
cylinder has to be known, since the volume occupied by a certain amount
of steam at a given pressure is a main element in the calculation of the
78 PETER KROES
power of a steam engine (p. 84). So, the first problem Pambour has to solve
is how the changes in the temperature of the steam affect the relation
between the pressure and volume . He argues that steam during its passage
through the engine is always saturated, i.e. is always in astate of maximum
density (p. 82):26
In general, therefore, the steam will always keep the highesl density corresponding to its
temperature during its operation in the cylinder; as if it had not ceased to be in contact with
the water from which it was produced .
This assumption enables Pambour to arrive at the required relation

between pressure and volume of the steam during its stay in the cylinder .
First, he derives a relation between specific volume (p. = V/S), temperature
and pressure which is also valid for saturated steam (pp . 68-69). Then he
observes that, for saturated steam, the pressure is a function of the
temperature. Elimination of the temperature leads to a direct relation
between pressure and specific volume (pp . 73 ff.). It is this relation that has
to be used for the calculation of the power of steam engines, rat her than
Boyle's law." Pambour shows that the use of Boyle's law may lead to
calculations of the work of steam engines which are 20% too high (p. 84).
In fact, Pambour introduces two different expressions for the relation
between specific volume and pressurer"
29254
I' = (engines with condensation)
1,784+p
30981
I' = (engines without condensation)
4,395+p
The first one is valid for low pressure engines, the second for high pressure
engines. Both relations are presented in the same form:
m
1'=--
n+p
with n and m empirical constants which are different for engines with or
without condensation.
Now it is easy to derive a formula for the relation between pressure and
volume for a fixed amount of saturated steam during the expansive phase
(p.83):
with MI (PI) and M (P) the initial and final volume (pressure)." This
relation, together with the principles of dynamic equilibrium and of
conservation of steam, is sufficient for solving the three problems stated
earlier.
Pambour first introduces the following notation (p . 176):
7r : the pressure of the steam at a certain moment after cut off,
A : the totallength of the path of the piston,
AI: the length of the path of the piston until cut off,
x the length of the path of the piston at the moment that the pressure of
the steam is 7r,
C C times a is the "c1earance" or the dead volume in the cylinder filled
each stroke with steam through which the piston does not sweep.
During the expansive phase, the differential of the work done, dW, is
equal to 7raox. Integrating over the whole expansive phase, adding the
work done before expansion, and equating the total amount of work done
to the load times the path, Pambour derives the following formula (p . 177):
A+C AI
aRA = a(AI + c)(n + Pl)[log nat - , - - + -,- - ] -nds: (6)
"I +C "I +C
This he calls the first general equation. The second basic equation follows
from the equality of the amount of steam generated in the boiler and
consumed by the engine (p . 179):
mS
(7)
n+P1
Together, these two equations give:
mS A+C AI
V= [log nat - , - + -,-] (8)
a(R + n) "I + C "I + C
This is the generalized version of Equation (4), describing the relation

between velocity, resistance and the evaporating capacity oj the boi/er.
The resistance R is composed of three elements (p , 180): (1) the net load,
denoted by r, (2) the friction of all the parts of the engine, which increases
80 PETER KROES
with the load r ; it is denoted by 4> + br with 4> the friction when there is no
load (r == 0); and (3) the back pressure p; for engines without conden-
sation, p is equal to the atmospheric pressure, and with condensation p is
the press ure in the condenser. 30
Putting
A+C AI
k == log nat - - + - - - (9)
AI+C AI+C
we get
mSk
v==--- (10)
a(n+R)
Pambour remarks that in this expression mS/(n + R) is nothing other than

the volume of the saturated steam of pressure R corresponding to the
amount of water S. Multiplication of this volume by k and division by the
surface of the piston a immediately gives the velocity v. This velocity does
not depend on the pressure P in the boiler, but only on the rate of
evaporation Sand the resistance R (p. 181).
The net power W is equal to (p. 183):
W == arv == mkS-av(n + p + 4» (11)

l+ö
From this expression it also follows that the net power W, just as the
velocity v, does not depend on the pressure P in the boiler, but on the
evaporation rate S [ibid.] . Neither does W depend upon the pressure PI in
the cylinder. This is, of course, a quite remarkable conclusion, since the
pressure in the cylinder usually had been one of the key elements in
computations of the power of steam engines.
Pambour goes on to determine the velocity for which the net power
reaches a maximum given afixed cut off, When designing steam engines,
it is important to know the maximum net power, because " a steam engine
may never be built so that its highest possible power is equal to the power
it has to deliver regularly, for otherwise no power would be left for an
accidental increase in the resistance" (p. 186) .31 Thus, under standard
conditions steam engines have to operate below maximum performance.
The maximum net power occurs for the lowest possible velocity; in that
situation the pressure in the cylinder PI is equal to the pressure in the

boiler P . The minimum velocity can be computed directly from the second
basic equation. The result is that the maximum net power W1 for a fixed
cut off is given by:
(12)
He observes that this maximum net power depends only on the

evaporation Sand the pressure P. To explain this, he adds (p. 192):32
Thus , the highest possible effect of the engine depends actually only on the evaporation Sand
the pressureP of the steam in the boiler ; that was to be expected, because the power of the
steam is determined only by these two . The cylinder and the motion of the piston, with their
dimensions, are just the means to transfer the power; they cannot increase or decrease it; and
also the speed of the motion ofthe piston cannot influence the highest possible effect, because
the speed can be increased and decreased at will only by the dimensions of the cylinder.
Thus , the construction of the boiler, not the construction of the engine,
determines the maximum net power!
Finally, Pambour discusses the "unbedingt-grössten Nutzwirkung" ,
i.e., the maximum net power for the most favourab/e cut off, This cut off
is given by:
(13)
For this value of the cut off, the net power is at a maximum. Pambour
rounds off his general theory of the steam engine by remarking that the
conditions for maximum net power differ from the conditions for the
maximum load that an engine can drive. In the following chapters he
elaborates this theory for different types of engines.
DESIGN AND THE TECHNOLOGICAL NATURE OF

PAMBOUR'S THEORY
According to Kerker, Pambour's Theorie de /a machine a vapeur has "a

fundamental and mathematical approach such as might be expected from
an applied physicist rather than a practical engineer" .33 Indeed, Pambour
82 PETER KROES
freely makes use of scientific laws and mathematical techniques to analyze

the operation of steam engines; his book presupposes that the reader is
familiar with mathematics and certain theories of physics. Nevertheless, he
does not produce a work of science. Although the book contains new and
important insights into the operation of steam engines, it does not
contribute to the theory of heat, or to the theory of mechanical power and
its relation to heat, nor to gas theory.34 In view of the explicitly stated aim
of the book, this is hardly surprising. Pambour sets out to solve practical
or technological and not scientific problems . In other words, Pambour's
book is a major contribution to the body of technological knowledge
(about steam engines), and not to scientific knowledge. In my opinion,
Pambour's theory offers some important eIues to an understanding of the
difference between both types of knowledge . It will be argued that this
difference is closely related to the role of design and design parameters.
Let us first dweIl a little longer on the aim of Pambour's theory. For
existing steam engines, his theory allows the computation of the velocity,
the load or evaporating capacity of the boiler, once two of these three
parameters are known . In particular, it allows adetermination of the
velocity of an engine, given the evaporating capacity ofthe boiler and load :
a problem, Pambour claims, engineers had been unable to solve before. It
also enables the engineer to calculate accurately the power of existing
engines and to determine the conditions for optimum performance
(optimum load, velocity and cut off) .
Pambour also intended his theory as an instrument for engineers for
solving problems in the context of designing steam engines. In his preface
to the German edition, Crelle writes (p. 1):35
The count de Pambour is the inventor of the theory of steam engines, that is, the mies by
which, with sufficient accuracy for its operation, the effects of an already existing steam
engine can be calculated from the evaporating capacity of its flue and from the pressure ofthe
steam in the boiler, and by which, conversely, when the engine still has to be built, the lauer
two quantities can be calculated given the required power.
And in the Introduction Pambour writes that the shortcomings of the old
theory were the cause of "many deceptions in the use of engines and of the
ensuing controversies between the buyers and manufacturers" (pp . 6_7) .36
Engines often turned out to be either too powerful or too weak. In the first
case, engines could not operate under conditions of maximum efficiency,
in the second, the pressure in the boiler had to be increased, frequently
resulting in explosions. Apparently, it was not possible to use the old

theory to design steam engines to satisfy the imposed specifications
accurately enough. Pambour intended his theory to remedy this situation.
It is instructive to compare Pambour's aim with Carnot's in his
Reflexions on the Motive Power 0/ Fire:"
The phenomenon ofthe production ofmotion by heat has not been treated from a sufficiently
general point of view. It has been treated almost exclusively with respect to engines whose
character and operation make it impossible for the full potential of the phenomenon to be
realized. In such engines, the production of motion is, as it were, curtailed and incomplete,
and it becomes difficult to perceive the principles underlying the process and to study its laws.
In order to grasp in a completely general way the principle governing the production of
mot ion by heat , it is necessary to consider the problem independently of any mechanism or
any particular working substance. Arguments have to be established that apply not only to
steam engines but also to any conceivable heat engine, whatever working substance is used
and whatever operations this working substance is made to perform.
Compared to Pambour, Carnot is clearly on a different track. He is

interested in the laws governing the production of motion (work) by heat
"independently of any mechanism" .38 The problem posed by Carnot is, in
my opinion, not itself a teehnological problem. Clearly, it is a problem
which arises from the technological context of steam engines; in that
context the question of how the input and output of steam engines are
related to each other is of central importance." But Carnot interprets this
question in such a way that it is separated from its concrete technological
setting. According to Krug.:" he creates "a model of the ideal technological
process at a level of abstraction at which construction parameters still play
no role" .41 This he does by characterizing the input and output of steam
engines in terms of physieal quantities. In other words, his intention is not
to develop a technological theory for a certain type of artifact, as Pambour
does, but he is primarily searching for a physical theory ab out the
conversion of heat in to work, in "any conceivable heat engine". Figure 1
clearly illustrates the difference between the two approaches. This figure
contains an elaborated and modified version of a classification of steam
engines discussed by Pambour (pp. 162 ff.) . The applicability of
Pambour's theory is limited by definite teehnological design eharaeteris-
tics, whereas Carnot's theory covers any kind of heat engine, that is any
kind of physical system converting heat into mechanical work.?
Nevertheless, the steam engine is omnipresent in the Reflexions" and
according to FOX 4 4 the book was primarily intended for steam engine
84 PETER KROES
engineers. And indeed, the content of Carnot's treatise is directIy relevant

to those engineers. The theoretical limit for the conversion of heat into
work , demonstrated by Carnot, is of course a valuable tool for evaluating
the efficiency of actual steam engines, and that is precisely what Carnot
does at the end of his book.?
Carnot's theory also contains significant rules for designing steam
engines. He derives the following principles for improving the perform-
ance of heat enginesr"
It is easy to irnagine a host of engines suitable for developing the mot ive power of heat through
the use of elastic fluids . But, whatever approach is adopted, we must not lose sight of the
following principles :
I. The temperature of the fluid must first be raised as high as possible, in order to secure a
large fall of caloric and thereby the production of a great amount of motive power .
2. For the same reason, the cooling must be carried as far as possible.
3. We must see that the passage ofthe elastic fluid from the highest to the lowest temperature
is brought about by an increase of volume. In other words, we must see to it that the
cooling of the gas is a spontaneous consequence of rarefaction.
These principles are important guidelines for designing steam engines. A

comparison of Carnot's design rules with those of Pambour conspicuously
reveals the different approaches underlying their theories. Pambour's
rules are derived from a theory about a specific dass of steam engines and
enable the engineer to determine the size of various parts of a steam engine.
Carnot's rules, on the other hand are derived from general physical
principles; they only concern thermal efficiency and are valid for any kind
of heat engine.:"
Let us now return to Pambour's theory. Pambour's focus on a
particular dass of artifacts is reflected in certain specific features of his
theory. These will now be studied in more detail. We will consider in turn
the domain of application of the theory, the basic principles of the theory,
and the kind of concepts employed by Pambour. A closer analysis of these
three aspects shows that the design of a particular dass of steam engines is
an integral part of Pambour's theory.
The Domain 01 Application
Whereas Carnot's theory about the transformation of heat into work may
be applied to any kind of heat engine, that is of any design whatsoever, the
PAMBOUR 'S THEORY OF THE STEAM ENGINE 85
Heat engines
Cornot
orr steam ether

engine engme engine
rotor reciproca- other

(turbines) ting piston Savery
Pombour
Fig. I. Classification scheme for heat engines.
design is a crucial element in Pambour's theory in respect to its domain of

application. His theory can be applied only to a restricted dass of heat
engines, namely piston operated engines using steam as their working
agent (see Fig. 1). Moreover, the validity of the principle of the con-
servation of steam requires the use of aseparate condenser (see below) .
Heat engines working without a piston, Savery' s steam pump and steam
86 PETER KROES
turbines, for instance, or those using another working agent, such as air
engines, fall outside its scope." The object of Pambour's theory is thus a
dass of artifacts based upon a given design, in which a reciprocating piston
is a key element. This design stillleaves room for all kinds of variations,
such as piston operated steam engines working under high or low pressure,
with or without cutting off and condensation (p. 164). Within the dass of
piston operated steam engines Pambour distinguishes nine different types.
His theory is intended to be a general theory which covers all these cases (p.
166).
A particular dass of artifacts (and not the physical processes occurring
in these artifacts, nor the physical theories describing these processes)
constitutes the main topic of Pambour's theory. For solving his three
principal problems, Pambour uses Newtonian mechanics and analyzes in
great detail the behaviour of steam in the cylinder on the basis of gas
theory. These theories by themselves, however, do not delimit the domain
of application of his theory in a significant way. Of course, Pambour has
to assurne the validity of the theories he uses; this is a conditio sine qua non
for the validity of his theory. The domain of application of Pambour's
theory is, however, further restricted in an essential way by certain
technological design characteristics. Beside the obvious "scientific"
constraints, imposed by the use of scientific theories , technological
constraints also confine the domain of application. These technological
constraints are closely related to the validity of one of the two basic
principles of the theory, namely the principle of the conservation of steam.
The Basic Principles 0/ the Theory
As we have seen, the foundation of Pambour's theory consists of two

principles: the principle of dynamical equilibrium and of conservation of
steam. The first principle goes back to the first law of Newton and is
considered to be valid in any context where forces and acceleration of mass
are involved. The use of this principle by Pambour is not related to any
specific features of the technological context of his analysis. In other
words, the validity of the application of Newton's law is not restricted in
any way by, nor based upon, the particular design of the heat engine under
consideration. Nevertheless, Pambour draws an extremely important
conclusion from this principle for the analysis of steam engines, namely
that the pressure in the cylinder is determined by the load and not by the
press ure in the boiler. Here, Pambour is exploiting scientific knowledge

for technological ends .
The situation with regard to the principle of the conservation of steam
is totally different. 49 It is not itself, like the principle of the conservation of
heat (caloric) or of energy, a scientific principle, nor is it a straightforward
consequence of applying physicallaws to a particular kind of system. This
principle is intimateiy related to the design and actual construction of
steam engines . The validity of this principle implies in the first place that
no steam is lost through leakage, through the safety-valve, for instance
(pp . 182-183). If steam is lost "because of the peculiar construction of an
engine";" then, Pambour remarks, his equations have to be adjusted.
Secondly, this principle entails that no condensation of steam may occur in
the steam engine. This means that all the parts of the engine with which the
steam will come into contact must have a temperature above the boiling
point of water, and that steam does not condense during expansion.
The conditions under which the principle of the conservation of steam
is valid can thus easily be expressed in terms of general physical boundary
conditions. If these boundary conditions are met, then the principle can be
justified on theoretical grounds. In other words, its validity can be
scientifically explained, So one might be tempted to consider Pambour's
use of this principle a consequence of applying gas theory to systems
satisfying these boundary conditions, and thus one might defend the thesis
that Pambour's theory of the steam engine is a piece of "applied science"
after all. However, explaining the validity of the principle of the con-
servation of steam in the case when certain general boundary conditions
are met is one thing; designing steam engines that actually satisfy this
principle is another. There is no logical, self-evident path from these
general boundary conditions to an adequate design . The design of the dass
of steam engines considered by Pambour is in no way an application of gas
theory or Newtonian mechanics, in the sense that it is possible to derive
from these theories a design satisfying the principle of the conservation of
steam; gas theory does not tell how these conditions can be met in designing
steam engines in practice.
The history of the steam engine shows that a whole series of techno-
logical innovations finally led to steam engines which more or less satisfied
the principle of the conservation of steam. The idea of using this principle
as a cornerstone for a theory of steam engines had been made possible by
a long development, in which the improvement of the performance of
steam engines by saving steam had been of central importance. The
88 PETER KROES
introduction of aseparate condenser by Watt, for instance, greatly

contributed to realizing the conditions under which the principle of the
conservation of steam may be supposed to be valid. Before Watt invented
the separate condenser, a lot of steam was lost in each stroke through
condensation on the inside surface of the cold cylinder. In 1769 Smeaton
conducted extensive experiments to determine how much steam was lost in
this way. He concluded that about three-quarters of the steam was
condensed and that only about one quarter was left for driving the piston.51
The invention of the separate condenser made it possible to keep the
cylinder and piston constantly at high temperature, thereby considerably
reducing the loss of steam. Without Watt's new design principle, the
conservation of steam could hardly have become a cornerstone for a
theory of steam engines. This shows the intimate relationship between this
principle and technological design principles . The assumption of the
conservation of steam may, therefore, be considered to be a technological
principle which is only valid for a restricted dass of artifacts.
In view of the foregoing, there exists a striking difference between the
computation of the power of a steam engine in Pambour's and Carnot's
theory. Given the relevant operating conditions of a steam engine, that is
the temperature of the boiler and the condenser, and the amount of heat
produced in the boiler, Carnot intends to calculate the (maximum) power
of an engine on the basis of a law of nature; for this, the design of the heat
engine is completely irrelevant. In Pambour's computation the design of a
particular kind of steam engines and its correlate, the principle of the
conservation of steam, playa decisive role; he would have been unable to
compute the work done by a steam engine solelyon the basis of Newtonian
mechanics and gas theory. His computation of the absolute highest net
power of a steam engine, therefore, cannot, from a physical point of view,
be placed on a par with Carnot's computation. Pambour's calculations
relate to a particular dass of artifacts and have no general physical validity,
unlike Carnot's computations. The fact that the absolute highest net
power is determined only by boiler capacity and pressure and is
independent of the size of piston and cylinder does not undermine this
statement; it does not lend any general validity of Pambour's result to all
kinds of heat engines. Pambour only shows that for the dass of piston
operated steam engines, the absolute highest net power does not depend on
the particular dimensions of cylinder and piston.
The Kind 01 Concept
As a consequence of the amalgamation of scientific and technological

principles in a single theory, technological concepts (design parameters)
and physical concepts appear side by side and are tied together in
Pambour's theory. In order to illustrate this, we shall have a closer look at
its two basic equations, (6) and (7).
They contain various kinds of quantities. Two of these, the empirical
constants n and m, are directly related to design characteristics of steam
engines, since n and m have different values for engines with or without
condensation. Pambour uses two different approximations for the
relation between the specific volume of saturated steam and its pressure,
one of which is valid for low pressures (i.e ., for engines with
condensation), the other for high pressures (i.e ., for engines without
condensation). But, clearly, the relation between specific volume and
pressure for saturated steam is intrinsically independent of any design
characteristic of a steam engine; this relation is a "law" of nature and in
principle it is possible to find a single expression for this law which is valid
for engines with and without condensation." As Pambour remarks, he
could have produced a single expression by using Biot's formula for the
relation between pressure and temperature for saturated steam (p. 73). But
that would have been mathematically difficult and the result would
probably be too complicated for practical use . Instead, he prefers to
present two simple expressions, each of which is valid for a different type
of steam engine.v
So the empirical constants n and m have no general physical status; their
significance can only be understood against the technological background
of Pambour's theory. Let us now turn to the quantities a, A and c. These
quantities concern the dimensions of certain parts of a steam engine and
are primarily technologically relevant. For a given engine they are fixed,
but in the context of designing an engine, they function as design
parameters. None of these design parameters have any general physical
significance in the sense that they are or can be embedded in physical
theories in which they are related to other physical quantities." The same
is also true for the design parameter AI; the value of this parameter,
however, can be varied for an engine at work.
The evaporating capacity, S, ofthe boiler is also a design parameter. It
is one of the central quantities of Pambour's theory, since it determines the
total power W of steam engines. S is therefore one of the main starting
90 PETER KROES
points in the process of designing a steam engine. Pambour does not

consider in detail the question ofhow S is related to the construction, or the
design parameters, of the boiler itself; he confines hirnself to abrief
description of the various types of boilers (pp . 101-107). It is interesting to
observe that Pambour does not devote much attention to the question of
how S is related to heat, that is how S as a design parameter can be related
to physical quantities, such as amount of heat, temperature and heat
capacity. He briefly describes Watt's law which states that the amount of
heat which is required to transform water into saturated steam is
independent of the pressure of the steam; in other words, the sum of
"latent" and "sensible" heat is constant (pp. 76-79). From this law and
Equation (11) he could immediately have drawn the conclusion that the
work produced by a steam engine is determined only by the amount of heat
produced in the boiler . He must have been aware of this, but apparently he
did not consider it worthwhile to point this out explicitly. Within the
context of Pambour's engineering theory, with its assumption ofthe con-
servation of steam, the evaporation rate S is a more significant parameter
than the amount of heat produced in the boiler: the evaporation rate is, for
instance, directly related to the speed of steam engines; the amount of heat
produced is not. Pambour, it may be concluded, did not deern it necessary
to translate this design parameter in terms of scientific concepts from the
theory ofheat, nor to study in closer detail the conversion ofheat into work
from a general physical point of view, as Carnot did. ss
Three quantities figuring in Equations (6) and (7) remain to be
discussed, namely the resistance or load R, the velocity v, and the pressure
in the cylinder PI . The first two belong to the three main (design)
parameters of Pambour's theory. From a technological point ofview they
are of paramount importance, because together they determine the actual
power delivered by a steam engine. The physical interpretation of these
design parameters poses no problems for Pambour. The resistance or load
is represented by a physical force; in doing this, Pambour is embedding
this design parameter into the physical theory of mechanical work. Note
that a characterization of the output of a steam engine in terms of physical
concepts had only become possible at the beginning of the 19th century.
By that time, technological notions that had been used for characterizing
the output or load of steam engines, like "horsepower" and "the number
of foot-pounds per minute" , also acquired a physical meaning. S6 This was
a crucial step which paved the way for a fruitful interaction between
science and steam power technology.S7
Finally, the pressure in the cylinder is an important technological

parameter within the context of the design and construction of cylinder
and piston. From a technological perspective, the cylinder pressure is
related to safety requirements, for instance, and plays a central part in
considerations about strength of materials. On the other hand, the
pressure in the cylinder is a physical quantity characterizing the state of
steam; as such its meaning is in no way related to the particular design of
an engine. The gas laws relate this pressure to the temperature and volume
of the steam. This situation allows Pambour to calculate the changes in
pressure in the cylinder, which he needs to know to determine the power
of an engine.
The synthesis of scientific and technological knowledge in Pambour's
theory is, in my opinion, made possible by this dual significance of
notions Iike "resistance" ("load") and "cylinder pressure" . In a techno-
logical context, these notions are treated like design parameters and are
part of a whole network of relations between concepts which are
technologically relevant. At the same time, these concepts are physical
quantities figuring in scientific theories which tie them to other physical
quantities. Note that not all parameters of Pambour's theory have such a
dual significance.
CONC LUSION
Reviewing the foregoing, it may be concluded that the design of a

particular type of artifact is an essential part of Pambour's theory. It is
not only a key element in demarcating the domain of application of the
theory, but it also provides the foundation for one of the two pillars on
which Pambour constructs his theory. Furthermore, a number of
concepts employed by Pambour derive their meaning directly from the
particular design involved. The possibility of translating some design
parameters into scientific concepts allows Pambour to exploit scientific
theories in solving technological problems. It is this combination of
design and scientific theories which makes Pambour's theory an
engineering theory as distinct from a scientific theory.
It is immediately obvious that in the case of Pambour's theory the
thesis that technology is applied science is utterly inappropriate. This
thesis implies that technological theories are derived from scientific
theories by inserting the relevant boundary conditions into these theories.
92 PETER KROES
This is not the way the principle of the conservation of steam, nor the two
basic equations are derived. In the case of the steam engine, the
technological design becomes from a physical point of view so compli-
cated that it cannot be translated into a set of boundary conditions which
can be fed into a physical theory. The design is no longer subordinated to
a physical theory by being simply its boundary conditions. Instead the
design itself becomes a crucial element in building a theory .
In order to avoid misunderstanding, it is necessary to set the above
distinction between scientific and technological theories briefly into a
broader perspective. The distinction put forward here does not imply that
a sharp boundary exists between scientific and technological knowledge.
As Mayr has remarked, any boundary between the two will be arbitrary;"
If we can make out boundaries at all between what we call science and technology, they are
usually arbitrary.(...)Trad itionally we regard physics as a science and the manufacture of
diesel engines as a technology, But what is thermodynamics, when textbooks are available
in all shades of emphasis, ranging from purely practical concerns to the most esoteric
theory?
Indeed, there seems to be a continuous spectrum of forms of knowledge

ranging from the purely technological to the purely scientific. This,
however, does not imply that it is meaningless to distinguish between
different types of knowledge; after all, we do distinguish between coloursl
It is easy to point out paradigmatic instances from both ends of the
spectrum as reference points."
Neither does the distinction entail that e fundamental difference exists
between scientific and technological knowledge with regard to the
manipulation of nature. Modern science, as it has evolved since the 16th
and 17th centuries , has proved to be an extremely fruitful partner for
technology . One specific aspect of modern science made this fruitful
partnership possible.?" This aspect is the experimental nature of modern
science. By adopting the experimental method, modern science became
ipso facto technologically relevant, for experimenting presupposes the
possibility of manipulating nature which in turn is the core of any
technological activity. Technological knowledge is by definition
knowledge about the manipulation of nature and of man-made devices.
Both types of knowledge are, therefore, deeply rooted in and spring from
the manipulation of nature. In that respect, there is no fundamental
distinction between the two.
The foregoing does not, however, exclude the existence of differences

"in both style and substance" between technological and scientific
knowledge due to diverging aims behind the manipulation of nature in
technology and science. According to Vincenti, these differences may
show up in "the way the knowledge is formulated". 61 The example of
Pambour's theory suggests that, at the theoretical level, engineering
knowledge is distinguished by the prominent place occupied by design
characteristics and design parameters.
Eindhoven University 01 Technology
NOTES
1 1 thank A. Sarlemijn, M. de Vries, E. Hornburg, G. Verbong and H . Lintsen for their
comments on an earlier version of this paper.

, Staudenmaier (1985), p. 107 ff.
J Staudenmaier (1985), p. 107.
• See Layton (1976), p. 695 and Vincenti (1982), p. 173.

• Cardwell (1971), p. 224 .
• Kerker (1960), p. 266 .
1 Kerker (1960). p. 266, where the reference to the quotation is given.
• Clausius (1876), pp. 247 ff.

• Singer (1985).
10 Dickinson (1938).
II Matschoss (1908).
" Hills (1989).

IJ Cardwell (1971).
" Kerker (1960) and Kerker (1974).

i s Kerker (1960), p. 268.
" Cardwell (1971), p. 224.

11 The following analy sis of Pambour's theory is based on the 1849 German translation by
Crelle of the French edition of 1844. Page numbers in the text refer to this edition.
" The invention of the Prony brake (around 1822) had to a great extent solved the problem
of directly measuring the power of a steam engine. The context of the invent ion of the " frein
dynarnometrique" is illustrative of the importance of finding reliable methods for
measuring the actual output of steam engines. Prony invented his method in connection with
a law suit about the efficiency of a Woolf engine for which he was asked to act as an expert
witness (see Payen (1976), p. 136).
" All quotations from Pambour's book are translated by the author; in each case, the
German text is given in a footnote . "Man glaubte zwar, im Voraus die Wirkungen einer zu
erbauende Maschine berechnen zu können, aber die Erfahrung zeigte, dass man der
verlangten Leistung immer nur dann gewiss sein konnt e, wenn die Maschine eine reine
94 PETER KROES
Nachbildung einer schon vorhandenen war, und dass die zu erwartende Wirkung durch jede
Abweichung ungewiss wurde ."
20 In spite of the great theoretical progress made during the nineteenth century in calculating
theoretically the power of steam engines, methods using empirically determined coefficients
of the kind rejected by Pambour were even being used in the twentieth century; see Ewing
(1926), p. 351.
21 " Es kommt auf Zweierlei bei einer Maschine an : auf den Widerstand, welcher sie in
Bewegung setzt, und auf die Geschwindigkeit, mit welcher dies geschieht. Zunächst also
ergeben sich folgende zwei Aufgaben.
Erstlich , Wenn für eine vorhandene Maschine die Geschwindigkeit bestimmt ist: den
Widerstand zu finden , den sie mit dieser Geschwindigkeit fortzutreiben vermag.
Zweitens. Wenn für eine vorhandene Maschine der Widerstand gegeben ist: die
Geschwindigkeit zu finden, mit welcher sie ihn in Bewegung setzen wird.
22 " Aber es gibt noch eine dritte Aufgabe, die aus den beiden vorigen von selbst hervorgeht,
nemlich:
Drittens. Wenn der Widerstand und die Geschwindigkeit gegeben sind : die Maasse der
Maschine zu finden, welche im Stande sein wird, diesen Widerstand mit der gegebenen
Geschwindigkeit zu überwinden. Bei den Dampfmaschinen ist, was hier gesucht wird, die
Grösse des Dampfkessels, oder , wenn man will, die der verlangten Wirkung gemässe
Verdampfungsfähigkeit.
23 In the German text the above quotation is immediately followed by aremark added by the
translator; it says: " Also, presumably, the dimensions of the cylinder and the other parts of
the engine" ("Auch wohl die Maasse der Dampfstiefel und der andern Theile der
Maschine.") For ereile the design problem therefore involves more than the capacity of the
boiler only . For Pambour, however, these other design problems, important as they might
be, are nevertheless subordinated to the most crucial design problem , namely finding the
capacity of the boiler given the desired outpur .
24 Pambour points out that the condition of the equality of motive force and resistance
concerns the mean value of both quantities during a whole stra ke, and not their
instantaneous values.
as "Folglich, so wie der Dampf aus dem Kessel in den Dampfstiefel tritt , ändert er seine
Spannung und nimmt diejenige an, welche dem Wiederstande des Kolbens gleich ist. Dieser
Umstand allein giebt der Theorie der Dampfmaschinen ihre Begründing und erklärt alles
N öthige."
2. "Im Allgemeinen wird also der Dampf während seiner Wirkung im Stiefel immer die
seiner Wärme entsprechende grösste Dichtigkeit behalten; eben so, als wenn er nicht
augehört hätte, mit dem ihn erzeugenden Wasser in Berührung zu sein."
This assumption was later abandoned by steam engine engineers; in the Rankine cycle " dry"
or superheated steam changes into "wet" or saturated steam dur ing expansion (see Kerker
(1960), pp . 266-267).
27 Boyle's law had been used for that purpose by Poncelet and Morin (Kerker (1960), p.
266).
28 The derivation of a single expression for the relation between specific volume and
pressure turns out to be rather complicated, partly for mathematical reasons (see p. 73).
29 Note that this expression is only valid when the amount of steam does not change .
30 For steam locomotives three other sources of resistance have to be taken into account,
namely the resistance of the air , the force necessary to set the engine in motion and the force
necessary for blowing the furnace (p. 180).
3 1 " Eine Dampfmaschine darf nie so gebaut werden , dass Das, was sie regelmässig zu leisten
hat, ihre möglich-grösste Wirkung sei, weil ihr son st für eine zufällige Vergrösserung des
Widerstandes keine Kraft übrig bleiben würde."
3 2 "Also hangt die m öglich-gr össte Wirkung der Maschine eigentlich nur allein von der
Verdampfung S und der Spannung P des Dampfs im Kessel ab; was auch natürlich ist, weil
sich danach allein die Kraft des Dampfs richtet. Der Stiefel und der Kolbenlauf, mit ihren
Maassen, sind nur die Mittel, die Kraft zu übertragen, ohne sie vergrössern oder schwächen
zu können; und auch die Geschwindigkeit der Bewegung des Kolbens kann auf die möglich-
grösste Wirkung keinen Einfluss haben , weil sie durch die Maasse des Stiefels allein sich
beliebig vergrössern und verkleinern lässt."
3 3 Kerker [19601, p. 266.
l4 In the historical introduction to his famous A Manual of the Steam Engine and Other
Prime Movers, Rankine remarks [18781, p. XXXI): "The investigations of the Count de
Pambour on the theory of the steam engine , although not involving the discovery of any
principle in thermodynamics properly speaking, were conducive to the progress of that
science by pointing out the proper mode of applying mechanical principles to the expansive
action of an e1astic fluid ."
ae "Der Herr Graf von Pambour ist der Erfinder der Theorie der Dampfmaschinen, das
heisst, der Regeln, nach welchen sich mit zureichender Genauigkeit für die Ausübung die
Wirkungen einer vorhandenen Dampfmaschine aus der Verdampfungskraft ihrer Esse und
aus der Spannung des Dampfs im Kessel berechnen lässt, und umgekehrt dieses beides für
die verlangte Wirkung, wenn die Maschine erst gebaut werden soll. "
38 " vielen Täuschungen bei dem Gebrauche der Maschinen und der dann folgenden
Streitigkeiten zwischen den Käufern und Verfertigern ."

37 Carnot [19861, pp . 63-64.
3. When discussing the possibility of steam engines with three cylinders , Carnot remarks
[19861, pp. 107-108: "We shall say no more on this subject , since it is not our aim here to
enter into the con structional details of steam engines ."
3 9 According to Fox , the background of Carnot's work is the debate ab out the economy of
the Woolf engine (Carnot , [19861, p. 8) : "It is true that Carnot constructed his theory
without reference to any particular engine, but I am convinced, none the less, that his work
should be seen as a contribution to the lively debate, cent ring on the que stion of economy
and conducted almost exclusively in France, that was sparked off by the recognition o f the
remarkable qualities of the Woolf engine in about 1815."
40 • Krug [19811, p . 8.
4 1 " ein Modell des technischen Idealprozesses auf einer Abstraktionsebene, in der
konstruktive Parameter noch keine Rolle spielen."
42 According to Herivel, an orientation towards generality and abstractness together with a
concern for applicability and utility was characteristic of the French mechanical school to
which Carnot belonged (Herivel [19761, pp. 89-92).
43 See Payen [19761, p. 125.
44 Carnot (1986), p. 2 .
4. See Carnot [19861, pp . 111-113 .

48 Carnot [1986], pp. 102-103.
96 PETER KROES
47 The last paragraph of Carnot's treatise c1early shows that he was aware that thermal
efficiency is only one factor alongside many others determining the overall efficiency of
steam engines (Carnot [1986)), p. 113.
4. Steam turbines were still at their infant stage at the time Pambour wrote his book; he
briefly mentions them in his survey of different kinds of steam engines (pp , 162 ff.).
49 Note that the principle of the conservation of steam has a counterpart in Carnot's theory;
Carnot assurnes that the amount of "elastic fluid" (air, for instance) does not change dur ing
a cycle. He analyzes a c/osed cycle (the same amount of working medium is used over and
over again) . This assumption, however, is not even explicitly stated as a principle, because
it is considered self evident in the context of Carnot's theory. Pambour, however, is
considering an open cycle in which in each cycle a fresh amount of working medium is
employed. For Pambour the principle of the conservation of steam is of central importance,
for it allows hirn to compute the speed of steam engines given the evaporating capacit y of the
boiler. Thus, both theories employ a similar conservation principle, but the status of these
principles is completely different.
50 "in Folge der eigenthümlichen Bauart einer Maschine" (p. 181).
.. See Pacey [I974], p. 136.

5 2 A good approximation of this law is given by the expression p V'6 115 = constant (Ewing
[1926)), p. 80.
5 3 Within engineering circles simple rules, which are only approximately valid, are often
preferred for practical reasons over complicated mies which are supposed to be exactly
valid.
54 There exists, for instance, no "theory of pistons" analogous to gas theory in which piston
characteristics, such as piston mass, surface and length of stroke, are related to each other .
55 According to Kerker [19601 , pp . 267-268 Pambour was not interested in the problem of
the efficiency of the conversion of heat into work , as was Carnot.

56 See Kroes [I991].
57 As Channell remarks [1982], p. 39: "So long as science and technology each had their own
independent framework, there could be Iittle direct Interaction between the two, In fact,
such an interaction requires the creation of a new body of knowledge - engineering science
-the purpose of which is to transform the concepts and discoveries in one area so that they
can be incorporated into the other area."
5. Mayr [1982], pp. 157-158.
5 9 General relativity and elementary particle physics are two reference points at the scientific
end; bodies of practical knowledge gathered in traditional crafts Iie at the other end of the
spectrum . Vincenti [1984]discusses a technological innovation in American airplanes which
came about without science.
60 Kroes [1989], p. 378.
6 1 Vincenti [1982], p. 137.

REFERENCES
Cardwell, D.S.L.: 1971, From Watt to Clausius, London: Heinemann .

Carnot, S.: 1986, Reflexions on the Motive Power of Fire (transl. & ed. by R. Fox),
Manchester: University Press .
Channell, D.F.: 1982, 'The Harmony of Theory and Practice: The Engineering Science of
W.J .M. Rankine' , Technology and Culture 23, pp . 39-52.
Clausius, R.: 1876, Die mechanische Wärmetheorie, Braunschweig: Friedrich Vieweg und
Sohn .
Dickinson, H.W.: 1938, A Short History of the Steam Engine, Babcock and Wilcox Ltd.,
Cambridge: University Press .
Ewing, J.A.: 1926, The Steam-Engine and Other Heat-Engines, Cambridge: University
Press.
Herivel, J. : 1976, 'Carnot and the French Scientific "Milieu" around 1824', in Sadi Carnot
et I'essor dela thermodynamique, Paris: Editions du centre national de la recherche
scientifique.
Hills, R.L. : 1989, Power from Steam, Cambridge: University Press .
Kerker, M.: 1960, 'Sadi Carnot and the Steam Engine Engineers', ISIS 51, pp. 257-270.
Kerker, M.: 1974, 'Pambour, Franccis Marie Guyonneau', in Gillispie, C .C . (ed.)
Dictionary of Scientific Biography, Vol. X , New York : Charles Scribner's Sons .
Kroes, P .A .: 1989, 'Philosophy of Science and the Technological Dimension of Science', in
Gavroglu, K. et al. (eds.) Imre Lakatos and Theories of Scientific Change, Dordrecht:
Kluwer Academic Publishers
Kroes, P .A. : 1991, 'Steam Engines and the Concept of Efficiency; Characteristics of
Technological Knowledge', Methodology and Science 24(2), pp . 79-97.
Krug, K.: 1981, 'Zur Herausbildung der technischen Thermodynamik am Beispiel der
wissenschaftlichen Schule van G.A . Zeuner', NTM-Schriftenr. Gesch. Naturwiss.,
Technik, Med., Leipzig 2, pp. 79-97 .
Layton, E.T. : 1976, ' American 1deologies of Science and Engineering', Technology and
Culture 17(4), pp. 688-701.
Matschoss, C.: 1980, Die Entwicklung der Dampfmachine: eine Geschichte der ortsfesten
Dampfmachine und der Lokomobile, der Schiffsmachine und Lokomotive, 2 vols.,
Berlin: Springer.
Mayr, 0 .: 1982, 'The Science-Technology Relationship' , in Barnes, B. and Edge, D. (eds.)
Science in Context, Cambridge (Mass .): MIT-Press.
Pacey, A .J.: 1974, 'Some Early Heat Engine Concepts and the Conservation of Heat', Brit.
J. Hist . Sc. 7(26), pp. 135-145.
Pambour, F.M .G. : 1894, Theorie der Dampfmaschinen (transl. A.L. Crelle), Berlin: G.
Reimer (originally published in French in 1844).
Payen, J.: 1976, 'La pratique des machines a vapeur au temps de Carnot', in Sadi Carnot
et I'essor de la thermodynamique, Paris: Editions du cent re national de la recherche
scientifique.
Rankine, W.J .M.: 1878, A Manual 0/ the Steam Engine and Other Prime Movers (ninth
ed.), London: CharIes Griffin and Co.
Singer, C. et al. (eds.): 1958, A History 0/ Technology, Vol , IV and V, Oxford: C1arendon
Press .
98 PETER KROES
Staudenmaier, J .M.: 1985, Technology 's Storytellers, Cambridge (Mass.): MIT-Press.

Vincenti, W.G. : 1982, 'Control-Volume Analysis: A Difference in Thinking between
Engineering and Physics', Technology and Cuttute 23, pp. 145-174.
Vincenti, W.G.: 1984, 'Technological Knowledge without Science: The Innovation of Flush
Riveting in American Airplanes, ca. 1930 - ca. 1950', Technology and Culture 25(3), pp.
540-576 .
ANORIES SARLEMIJN ANO MARC OE VRIES·
THE PIECEMEAL RATIONALITY OF

APPLICATION-ORIENTED RESEARCH
An Analysis 0/ the R&D-history Leading to the Invention 0/ the
Plumbicon in the Phi/ips Research Laboratories
INTROOUCTION
The Plumbicon is a television camera pickup tube which was developed by

the Philips Research Laboratories. I It was made public in 1962 and even
today is considered the ultimate in pickup tube designs. This was
confirmed by G.O . Towler of the Department of Trade and Industry of
the UK during the International Conference on the History of Television':
The Plumbicon has been so successful for broadcasting applications , that no strong
competitor has yet emerged to replace it.
Later designs , like the Japanese saticon, are not regarded as real
improvements; the Plumbicon can be considered as the successful closure
of a long sequence of pickup tube designs, going back into the late 19th
century:
designs 0/ tubes with mechanical pickup devices:

1884 Germ : Nipkov disk
1924 UK : Baird system
1925 USA : Jenkins system
designs 0/ tubes with electrica//electronic pickup devices:

1932 UK : emitron (super emitron, 1934)
1933 USA: iconoscope (image iconoscope, 1934)
1937 USA : orthicon (irnage orthicon, 1946)
1947 Fr eriscope
1950 USA : vidicon
1962 NL Plumbicon
In our analysis ofthis sequence of designs, we will pay special attention to
various factors that have influenced the development of television camera
pickup tubes. This may help us to gain a better insight into the way
sequences of designs develop . In general we may distinguish the following
types of influential factors .
99
P. Kroes and M . Bakker (eds.), Technological Development and Science in the lndustrial
Age,99-131.
100 ANDRIES SARLEMI JN AND MARe DE VRIES
(1) S-factors ("scientific factors") are phenomena that are studied in

exact sciences and havecharacteristics that justify the expectation that
they can be manipulated in the context of a technological system.
(2) T-factors ("technological factors") are factors that can be described
by a technological model.
(3) M-factors ("market factors") cause the so-caIled "market pull"; these
factors have a concrete content for existing products in a certain
situation. An example is the citizens' need to be charged for a product
with a predictable price. Without production costs and the income
expectations of population groups cost/benefit analyses will become
speculative . And yet, the M-factors exert an influence even in this
speculative form: in that form they have led to many innovations. The
R&D efforts to achieve television technology form a clear example .
(4) PI1-factors ("political and/or juridical factors") have a political
and/or (quasi-) juridical nature, although they are directly related to
production systems. To these factors belong agreements on line
systems for television, the wavelengths for broadcasting and
regulations like the American ASA, Dutch NEN and the German DIN
standards.' Such regulations, of course, have a prescriptive character.
Industrial and national interests play a role in their establishment.
The g, and T-factors together form, for instance, the object of the techno-
logical know-how developed in the Philips Lab and have been applied to
the production of the Plumbicon since 1962. However, without the
required M-factors and sufficient regulations - the P/J factors, for
instance with respect to the line system - mass production is not possible.
It will be clear that the involvement of this variety of factors pro duces
great problems in the context of planning application-oriented research.
This concept of "application-oriented research" is slightly problematic
because it has given rise to contrary opinions.
- Some think that inventions are the result of serendipity: of unplanned
coincidences that individual researchers experience." This would imply
that (sequences of) designs or innovations can not be planned.
- Others think that industrialists are smart enough to do research only
when risks are excluded by prior cost/benefit analyses . This would
imply that fiascos have to be ascribed to bad management and that
successes are the result of weIl planned activities .'
In the field of electronics, to which the pickup tubes belong, the truth
probably lies somewhere in the middle. Innovations in that area depend
on a variety of factors and at least some of these factors are the object of
APPLICA TION-ORIENTED RESEARCH 101
guesses. In other words: these innovations are the results of global

planning with uncertainty about influential factors . The analysis of the
four factors mentioned above in the context of the "chain" of designs that
led to the Plumbicon will provide a meaning for our concept of
"piecemeal rationality in design" ." It will become evident how the Philips
corporation negotiated all these factors and thereby tried to reduce its
risks. This process is best described as one of piecemeal rationality instead
of complete serendipity or strict rational planning.
Piecemeal rationality is an aspect of the sequence of designs: every
design is an attempt to realise an optimum reaction to the S-, T-, M- and
P/J-factors of the moment. In the case at hand, these factors at the end
ofthe 1920s and the early 1930s allowed only for pickup tube designs that
cou ld be produced in the context of small industries like the Baird
Company and not in that of large companies such as Philips. By the end
of the 1930s the S- and T-factors were changing and became more
favourable to mass production. Gilles Holst, the director of the Philips
Lab, followed this evolution attentively and reacted by enlarging the
television research group. He was still in doubt about the M-factors.
Therefore, he maintained a research group working on an alternative (the
so-called "house cineac ") . But in 1947 the M-factors were c1early in
favour of the commercial mass production of television sets. As a
"logical" consequence of this new situation the "alternative" research
was stopped and all attention was fixed on television and on the
Plumbicon research . Dur overall conclusion is that the sequence of pickup
tube designs shows a rationality of a piecemeal character:
- strict rational planning of the final success, the Plumbicon, can be
reconstructed by analyzing the transition from its predecessor, the
vidicon (a single, well-defined program, beginning with Nipkov and
leading to the Plumbicon, cannot be verified);
- every transition has its own rationality;
- step by step, an optimum reaction on the different factors can be
found;
- in the whole sequence of designs, however, only some steps of global
rational planning (and not a strict rationally planned program) can be
discovered; a symptomof this aspect is that the work on alternatives
could not yet be excluded at the end of the 1930s.
A more general remark about the kind of innovation we are analyzing
should be made here. This is an innovation in the field of microtechno-
logy , Dur conclusions, therefore, will be restricted to this field . Micro-
102 ANDRIES SARLEMIJN AND MARC DE VRIES
technological innovations have a more complex character than those in

experience technology like, for instance, a new corkscrew, or in macro-
technology, such as the aerodynamic aspects of a new aircraft.? The
difference between these technologies is caused by the S- and T-factors:
in experience technology the basic principles are empiricallaws borrowed
from the natural sciences. In macrotechnologies fundamental theories
serve as "calculation methods":" their principles are not discussed,
nor are a subject of doubt in R&D developments . In microtechnologies,
on the other hand, basic principles are often still a subject of
fundamental research : here, innovations often require special R&D
investments.
We do not believe that the usual innovation literature is correct in
treating corkscrews, airplanes, transistors and chips indiscriminately as
innovations of one and the same kind. The differences in know-how
leading to these different types of innovations are too fundamental to
justify such a simplification. In our example, of course, most attention
must be paid to S- and T-factors, but we will not ignore decisive influences
of M- and PI1-factors .
I. NIPKOV'S FIRST IDEA ABOUT TELEVISION RECORDINGS (1884)
The television was introduced without any preceding analogue. At best

one might think of the cinema, but the difference between cinema and
television is so crucial that the analogy in terms of M- and P/J-factors is
quite weak. For that reason it was extremely difficult to imagine, even in
the 1940s, a mass production of television and a corresponding market.
And yet the concept of television is old for it sterns from the 1880s. In
1884 a German student, Paul Gottlieb Nipkov (1860-1940), invented a
design which is shown in Fig. 1. According to Nipkov broadcasting pro-
ceeded in three steps :"
(1) The image B to be broadcast is illuminated via a lens K and the so-
called Nipkov disk N (see Fig. 1). This rotating disk divides the image
into a lattice of image points . These points are illuminated one after
the other. We can see the result of this in Fig. 2: the scanning of the
image starts in the upper left-hand corner and follows the arc lines
from left to right.
(2) The reflection of the light is then captured by a selenium cell S. Its
intensity varies with the parts of the image to be broadcast. As a result
APPLlCATION-ORIENTED RESEARCH 103
L
"
Fig. 10 '0 The Nipkov disk ; S is the selenium cell and i is the current of this cell.
the current i in the cell also varies: it will be strong in the lighter
parts and weak in the darker parts of the image.
(3) Corresponding to the current a lamp will emit more or less light
when reproducing the image. If the same type of disk is used in
recording and reproducing, the image will reappear.
Yet the realization of Nipkov's idea was impossible at the time he
developed it. There were four problems ; for at least three of them the S-
and T-factors required to solve them were lacking in Nipkov's time:
Fig. 20 The image of two wornen, recorded and displayed with a Nipkov disk .
(1') In reproducing the image the brightness and contrast depend on the
speed with which a lamp can follow the variations in current. Due to
current variations there were indeed variations in light intensity, but
the reaction of the lamp was too slow: once heated, the filament kept
on emitting light for a relatively long time . In that sense television
presupposes a microtechnological development like electronics.
(2') The second problem is the slow reaction of the selenium cell. Im-
provement of this cell presupposes an increase of knowledge of the
relevant materials.
(3') In the third place no devices were then available for amplifying the
signal from the cell. This also presupposes electronics.
(4') A fourth problem was inherent in the mechanical approach. Along
the border of the Nipkov disk 625 holes (depending on the line system
chosen) must be made. One can calculate, that because of the size of
those holes, the disk should have a diameter of 5 meters for an image
size of 24 millimetres! The disk should turn at a speed of 25 rotations
per second . This rotation, in addition, must be free of vibration,
which is no easy design requirement to fulfil.
Therefore, television was both an ingenious and an impossible idea of a 24
year old German student. But at the age of 77 Nipkov could admire the
realization of his idea : he was present at the demonstration of television
at the Berliner Funkausstellung of 1937.
2. THE FIRST COMMERCIAL TELEVISION OF THE BAIRD

COMPANY AND THE FIRST INVOLVEMENT OF THE PHILIPS LAB
The idea of television matured with the emergence of electron theory and
electronics . Crucial developments in this respect were:
- the discovery of the electron in 1895 by Perrin and Thomson,
- the design of the cathode ray tube in 1897 by Braun, and
- the first design of a rectifier/detector (by Fleming in 1902) and that of
an amplifier (by Lee de Forest in 1906).
The first person to recognize the importance of these developments for
television was A.A. Campbell Swinton. His publications between 1909
and 1924gave an insight into the essential elements of television based on
electronics but he was unable to put his ideas into practice." For the first
applications of electronics we have to turn to the work of J.L. Baird in the
UK and of C.F. Jenkins in the USA.
APPLICA T10N-ORIENTED RESEARCH 105
Baird became the pioneer in Britain because of his use of triode

amplifiers and the replacement of the selenium cell by agas filled
potassium photocell." The Nipkov disk that Baird used had a grid of 30
verticallines and a speed of just 5 images per second (which of course does
not yield a good reproduction of moving pictures). In January 1926 he
showed his results to the Royal Institution in London by transmitting the
static image of the shadow of a Maltese Cross. In 1928 he was the first to
seIl three models of television sets (see Fig. 3).
" T ELEVISOR ..
ORDE R FORM .
7.
U. II:U T,"LL'I~'O" DC'·CI.O".,(.. r <.:0 .11'.. .. ' . Ll loIlTLI.l.
19Z8 IU. Lo "c. AC_ I
u.,...""... wC:
"1 0'<01.'l.Urr' l 10 mc Morlt' n ' 1OC'C .liCII In )_, l?lS Llll l. ,,=u,
.1 ,I.. "'I.' I'.. a .... l . •"tl /CCl ro J" .I ur-' lhe' CO~dl lo,)o1 " rn4or-e>.1
I... u ".
S., ,
I
--
-::.., r-n
1.',. ( /..Jl
PtHl,. 1 A-'J .CJI
1.4 110 TlU'YI$ION DfVU O~I."'"

CO MI" ANY. LTD
u l \Ol'OO .. e- L ~ , W L.l
",.
Fig, 3. The front page of the Baird Company's brochure for television sets and the
order form for the sets.
The selenium cell had also been replaced by Jenkins in the USA; he used
an alkali metal photocell. In June 1925 Jenkins had demonstrated his
system with bevel-edged glass disks and transmitted the moving image of
a slowly rotating model windmill ." This, however, was still a far cry from
modern television . From the contemporary point of view we can never-
theless understand the enthusiasm with which these results were received
and the amazement they evoked. In the Philips Lab, too, a discussion
sprang up as to whether Philips ought to do something in this direction as
weIl. The result was that Philips did indeed enter the field.
In September 1927 M.J . Druyvesteyn was employed by the Philips Lab
106 ANDRIES SARLEMIJN AND MARC OE VRIES
and was given the assignment by its director Gilles Holst to "do something
on television"." Only a few weeks later, on 4 October, a patent was
registered in Holst's name ." In December of that year Druyvesteyn gave
his first demonstrations: the first time to the Dutch Physics Association
and then - for two weeks - to a broader audience. The receiver was just
a few rooms away from the transmitter. For broadcasting the short wave
was used (7.85 m according to Van der Pol). Druyvesteyn's pickup and
receiver devices were based on the use of a Nipkov disk, a photoceIl and
a neon incandescent lamp . At first a static image of a lantern was
transmitted; later a film was used."
Neither Druyvesteyn hirnself, nor the board of the Philips Lab
expected much success from work in this field. That is why it was
abandoned shortly afterwards (in 1929 or 1930). Some years later the
work was resumed, but by that time a number of new developments had
gathered speed abroad.
3 . THE ICONOSCOPE (1933) AND THE IMAGE ICONOSCOPE (1934);

THE PHILIPS LAB lOINS THE FIELD ONCE MORE
In the history of pickup tubes the work of the Russian emigrant Vladimir
Kosma Zworykin (1889-1982) was very influential: he replaced the
Nipkov disk by an electronic component, which was a crucial step in the
development. By substituting the slow mechanical scanning by the much
faster electronic scanning, it became possible to split up the image into
many more lines, while keeping the number of images per second
unchanged. The second reason for the importance of this step was the fact
that the image was now stored during the time that elapsed between two
scans, which produced a much higher tube sensitivity. In fact the whole of
this further development was aimed at making optimum use of this
intermediate time in order to improve the light sensitivity of the tube.
In 1925 Zworykin applied for a patent for this concept. He then
worked at the Westinghouse Electric Corporation (WEC) but he was
unable to convince his superiors of the benefits of television. In 1929 a
Radio Corporation of America (RCA) official visited the WEC Lab and
became acquainted with Zworykin's results. He became enthusiastic and
offered to Zworykin a position as director of the RCA Lab. Zworykin
accepted and in 1933 he could write that his design had been put into
practice and functioned weIl: the iconoscope."
APPLICATION-ORIENTED RESEARCH 107
As stated before, the primary important feature of the iconoscope was

the electronic scanning. Part ofthe iconoscope was a modified Braun tube
(Fig . 4). The three electrodes - familiar from the triode tube (the cathode
K with filament F, grid G and anode A) - were transformed into an
electron gun that could be directed towards a chosen position on the
screen S by the focusing electrode E . Another vital function was
performed by the pairs of deflection plates Hand V: they steered the beam
horizontally and vertically. An increase or decrease of the voltage on the
plates allowed the beam to scan the image to be transmitted.
V H
G E A
K
· ·················· ················v
F
'--_....:.-_---+-+-.........
Fig. 4. The modified Braun tube.
The second important feature was the electrical storage of the image. The
scanning of the image was not done immediately, as in the Nipkov system,
but the image was first recorded by a device consisting of three layers (see
Fig .5):
- a mosaic layer, C, of cesium grains that are isolated from each other,
- an isolating mica layer, M, and
- a conducting metal signal plate, P .
Through the lens L the image falls on the mosaic of cesium grains.
Depending on the amount of light in the various parts of the original
image, a number of electrons are emitted by every grain and are captured
by the collector anode Ac . Therefore, the grains become positively
charged, depending on the light intensity. This positive charge is stored
because the grains perform a dual role:
- on the one hand they transform light rays into electric energy, just like
photocells,
108 ANDR1ES SARLEMIJN AND MARe DE VR1ES
- on the other hand each cesium grain, together with M and the con-
ductor P, behaves like a capacitor that stores the electric energy; the
grain remains positively charged until it is discharged by the electron
beam.
Next, the scanning beam discharges the grains one after the other. When
a grain is discharged this causes a small current. This discharge current
produces a difference in voltage over resistor R, which is sent to amplifier
Am. This is now called the video signal, which varies according to the
amount of light in the image.
The iconoscope had one serious disadvantage. When the target was hit
by the scanning beam , secondary electron emission occurred. Some of
/ I \ R
C M P
Ac C M P
I /
Fig. 5. The 1933 iconoscope.

these so-called "secondary electrons" went to the collector anode Ac,

others fell back onto the cesium layer. This "redistribution" of electrons
is hard to control: the electrons are irregularly divided over the grains. As
a result the image becomes spotted. The struggle for control over this
secondary emission was to be a crucial issue in the design of pickup tubes
for a long period .
A successful attempt to improve the sensitivity of the pickup tube is the
design ofthe image iconoscope patented by Zworykin in 1934. In this tube
the picture was not stored immediately in the cesium grains; the light first
hits a photocathode Pc (Fig. 6). By using a continuous photocathode
instead of a mosaic, the surface of the granular target was used more
efficiently and the light sensitivity of the tube was thus improved. When
the photocathode was illuminated, photoelectrons were emitted . These
electrons, directed by electromagnetic fields of the focusing coils, Cf,
went to the target, which consisted of a continuous mica layer, M, that
had been exposed to cesium vapour to make it electron emissive, and a
conducting signal plate, P . The electrons hit the mica layer at high speed,
thereby causing secondary electrons to be emitted from the target. A
collector anode , Ac, at a high voltage (1000 volts) caught the electrons and
prevented them from returning to the target.
The Philips Technical Review appeared for the first time in 1936.
During its first year two articles were published by Van der Mark on
television experiments in the Philips Lab . The first article dealt with the
normal iconoscope. From the second article it appears that there was an
interest in the transmitter problems, They were related to the flickerings
of the image that are caused by the mutual interaction between lines: a line
that has just been scanned influences its neighbour. According to Van der
Mark this could be corrected by increasing the number of lines per second,
but he adds a warning:
This, however, would mean that the maximum modulation frequency is doubled, which
should be avoided."
Therefore, a different solution had to be found and Van der Mark

suggested interlinear scanning. This is the alternating scanning of odd and
even lines. But the use of a modulating frequency, however, would cause
problems with regard to the number of transmitters for television broad-
casting.
110 ANDRIES SARLEMIJN AND MARe OE VRIES
Cf
I I --------+-----~
1000 V
Cd=deflection coifs
Fig.6. The 1934 image iconoscope .
4 . DIFFERENT DESIGN TRADITIONS
4.a. The Emitron, the Eriscope, Farnsworth's Tube, and the Orthicon:
How S-, T- and PIl-factors Separated the USA and the UKfrom the
European Continent after the War
More or less independently of Zworykin, other developments went on in

different countries . A second development took place in the UK where the
emitron was designed by EMI in 1932. The emitron principle is roughly
the same as that of the iconoscope . Its improvements led to the super
emitron, which was taken into service by the BBC. Isaac Shoenberg, EMI
Director of Research, took the courageous decision to go for a 405 line,
50 Herz interlaced system that gave the UK the first "high definition"
public television service in the world ." After the Second World War
England introduced the cathode potential stabilized emitron (CPS
emitron), analogous to the American orthicon (see below).
As a third route, after Zworykin and EMI, we should mention the
French eriscope of 1947. This tube was based on a different line standard-
ization."
Then there was Farnsworth's tube in the USA . In his design the image
was projected on a photocathode plate, Pe, and immediately scanned,

without storing the image. As the signal/noise ratio was insufficient the
tube could only be used in the case of extremely high light intensity. 21 And
yet Farnsworth's idea was not completely forgotten. Part of his idea was
the release of an extremely large emission of secondary electrons by a
secondary emission amplifier S (see Fig. 7). This was realized through a
number of prepared layers (Al, etc.) where one electron, via a number of
layers, can release hundreds of millions of other electrons . This enabled
Farnsworth to obtain a very high image clarity. We will come across this
part of his device in later designs such as the image orthicon .
Cf
... ... ... ... Pe

... ... ... ...
.------
.. ...- ... ...... ...... ...-... ...... ......
...
....
... ... ... ...
+1 Cd Cf
Fig.7. Farnsworth's tube .
An initially more promising route led to the design of the orthicon. In the
USA this tube was developed by H. Iams and A. Rose in the RCA Lab . 2 2
It was developed further during the war, resulting in 1946 in the image
orthicon.
The word "orthicon" sterns from the Greek "orthos", which means
"perpendicular" . In contrast to the iconoscope in the orthicon the
electron beam hits the target perpendicularly. A double sided target was
used in this tube: in Fig. 8 the beam from the electron gun (cathode K and
anode A) scans the target F at the "back" side (seen from the image side).
Fis a semitransparent plate. The lens L projects the image on its "front"
side. F is scanned with a low velocity beam: the velocity with which the
112 ANDRIES SARLEMIJN AND MARe DE VRIES
electrons hit the target is low and secondary emission can thus be better
controlled. In the iconoscope there was a large potential difference
between the cathode ofthe electron gun and the target (about 1000volts);
this caused a high velocity scanning beam . The voltage of the anode was
about the same as the voltage of the target. The iconoscope was therefore
called "anode stabilised" .
In the orthicon, however, the gun cathode potential is the same as the
target potential (the orthicon is therefore called "cathode stabilised") and
there is a large potential difference between target and anode . This caused
the electrons coming out of the target to go to the collector anode Ac that
is located between electron gun and target (in the figure this is the
"returning beam" G) and then to fall on the collector Ac. Since the
invention of the orthicon low velocity beam scanning became the main
trend in the Anglo-Saxon pickup tube designs. Its potential for outdoor
recordings of acceptable quality was recognized . For indoor recordings,
too, less light was needed because of the better light sensitivity of the tube .
Attempts to further improve the light sensitivity of the orthicon were
made and these led to the image orthicon developed by the RCA Lab
during the war. The image orthicon is drawn schematically in Fig. 9. The
main difference from the orthicon is the added photocathode Pc and the
electron multiplier S. which we have already seen in Farnsworth's design.
A
c
~ t~·-~~---·
I - - - .' - -
F
Am
100 V
R
Fig. 8. The orthicon .

APPLICATION -ORIENTED RESEARCH 113
Like the orthicon, it had a scanning beam that went back and forth. On
its way back the beam passed through a hole in the anode and hit the
electron multiplier S. A double sided target was used, which was also
borrowed from the orthicon. Thus the image orthicon combined many of
the principles which were already available, such as low velocity beam
scanning, a double sided target, the combination of a photocathode and
a target, and the electron multiplier .
Cf Cd
II r..~::::=====;:::,
Fig. 9. The image orthicon.
In 1951 Kerkhof and Werner , who were in charge of Philips' television

development, were convinced that, theoretically, the image orthicon was
the optimum design. Its light sensitivity was a thousand times better than
the iconoscope ;" Initially, simplicity had been the leading design
principle of the orthicon, but corrections and modifications had now led
to the image orthicon, which was a very complex device. In the context of
design, device complexity is to be avoided because it does not allow low-
cost production and it increases the danger of malfunction. Therefore, the
image orthicon could not be the final step (on this point Kerkhof and
Werner were wrong). During further development the subsequent P/J-
factors in a certain sense separated the design traditions. The development
mentioned above led to a contrast between the USA and the UK on the
one hand and the European Continent on the other hand. Even before the
war the English had standardized their images at 405 lines as we have seen.
The Americans stuck to 525 lines. In 1950 Belgium, Denmark, Italy, the
Netherlands, Sweden and Switzerland decided to use a 625 line system.
France wanted a better quality and used 819 lines.
114 ANORIES SARLEMIJN ANO MARe OE VRIES
These agreements and disagreements are not free of national interests:

each country fostered its own design. Tubes with a low velocity scanning
beam, like the orthicon, the CPS emitron and the image orthicon
prevailed in the USA and the UK. The image iconoscope became a
specifically continental affair. 24 A relationship can be observed between
the number of lines and the scanning beam electron velocity. Low velocity
scanning electron beams can be focused better. Therefore a large number
of lines per image needs a low velocity scanning beam . Thus different
choices in the number of lines per image led to different tube designs in the
Anglo-Saxon countries and on the Continent.
4.b. The Vidicon (1950), the Storting Point of the Work on the
Plumbicon as the Unification of the Different Design Traditions
For the vidicon, which is based on the use of photoconduction, S-, T-, M-,
and P/J-factors can be indicated, that were mainly present in the USA.
Especially there, both during and after the war, the technological
know-how with respect to semiconductors had increased because of the
work on radar and the transistor. This had caused a growing interest in
solid state and surface physics, which in turn influenced the development
of television. The vidicon design of the RCA Lab was based on a better
understanding of the principles of semiconductors and conductors. In
Europe there was skepsis concerning this design: it was doubted whether
semiconductors would be able to yield stable pickup tubes. 25
After the Second World War specific M-factors in the USA stimulated
the development of television: they were mainly related to the standard of
living and the possibilities for spending money on luxuries, such as
television sets. In postwar Europe, until the early 1950s, there was no
prospect for such markets."
The M-factors were governed by the P/ J-factors, of which the varieties
in line systems are an outstanding example. And these systems
corresponded to certain designs for pickup tubes, as we showed above .
Therefore the vidicon cannot be said to have originated from serendipity
or happy coincidence: it was based on the deliberate assumption that it
was possible to combine and exploit those factors.
That was the concept behind the work in the RCA Lab . P.K . Weimer,
S.V. Forgue and R.B. Goodrich worked on the vidicon and in 1950
presented their new invention in the March issue of Electronics and later
APPLICATlON -ORIENTED RESEARCH 115
in the RCA Review. This background of technological and social embed-

ding will make it easier to understand the design of the vidicon. In
particular it will become evident why the vidicon deviated from the four
routes mentioned in Section 4.a (iconoscope, emitron, Farnsworth's tube
and orthicon). Semiconductors allow electron emission to be replaced by
a completely different mechanism, viz., photoconduction.
All four routes considered so far used electron emission: the scanning
beam throws out electrons from the target. Characteristic of this line of
development is - as we have seen - the struggle with the negative side
effects of this secondary emission. Semiconductors enabled the RCA
people to avoid those problems and to use a different principle:
photoconduction. This means that the conductivity of a target depends on
the amount of light that falls on the target.
This phenomenon was discovered as far back as 1873, when Joseph
May checked the Transatlantic telegraph cable: there were fluctuations in
the current of his measuring instrument that seemed to be related to the
light intensity on a selenium rod in his instrument. 27 He sought an
explanation, but to no avail. Today we know that an explanation could
not have been given before the development of solid state physics. And
yet, early in the development of pickup tubes, attempts were already made
to exploit this effect. For instance in 1925 C.A. Sabbah applied for three
patents for pickup tubes , one of which used photoconductivity. This was
probably never built. Efforts in the 1930s and 1940s did not yield
successful results. The principle of electron emission therefore remained
dominant until the vidicon was developed .
Figure 10 shows the design scheme of the vidicon. The upper part of
Fig. 10 (which gives parts Si and Se in more detail) shows its most
important innovation, the target, with:
- a conductive signal plate Si and
- a photoconductive semiconductor plate Se (of selenium or antimony
sulfide).
The conductive signal plate Si faces the light image, but as it is trans-
parent, the light passes through it and hits the photoconductive layer Se.
When there is no light the voltage of this layer Se is driven towards the
voltage of the cathode by the scanning beam that periodically hits it. The
signal plate is connected to the earth, so that, in the dark, there is a
potential difference between Se and Si. No electrons can pass this voltage
difference , because Se does not conduct in the dark . When light hits the
layer Se, however, it becomes conductive and electrons can go from the
Cd Cf
Fig. 10. The vidicon.
photoconductive layer Se to the signal plate Si, whereby the voltage of Se

becomes locally higher. When the scanning beam hits this illuminated
place with its somewhat higher voltage, the original potential is restored,
which causes a small current that is sent to the amplifier as the video
signal.
A problem was that the photoconductivity process with the chosen
target material was too slow to follow rapid changes in the image. That
made the image vague when it changed rapidly, just like photos that have
been made with too slow a shutter speed. In the RCA Lab this
phenomenon was accepted as an inherent feature of this process, while at
Philips it was believed to be possible to choose other materials that would
yield better results. These contrasting ideas gave rise to different
management decisions by RCA and Philips in the Fifties. Even at that
time it was evident that, as soon as there was a better version of the
vidicon, the image orthicon would be driven out of the market. A second
disadvantage of the vidicon exacerbated the problem: at low light
intensities there was still a small signal current (the so-called "dark
current"). Given these disadvantages one can understand the techno-

logical goals set by R&D managers: they tried to find suitable materials
that would respond better to fast changes in the image and would
eliminate dark currents.
It was the Plumbicon in which these goals were realized." The
Plumbicon can be regarded as bridging the gap between the two design
traditions: the Anglo-Saxon and the conti nental one. It can be seen as the
result of fruitful research into suitable semiconductors both by chemists
and physicists in the Philips Lab in the 1950s.
It was not only the know-how of scientists that was involved here:
there was also an intense cooperation between the research lab and the
production side. The initial production was hampered by problems of
bursting glass." To eliminate this, the skills of glassblowers were needed.
Pickup tubes are masterpieces of glassblowing skill and because of the mot ivation of very
skilled glassblowers the most unorthodox constructions have been realised. Mr . Scham pers
was the man with the golden hands , who began as a glassblower , but by self study rose to
the level of scientific assistant , and has contributed significantl y in all stages of Plumbicon
development. He invented the completely new vacuum melting connection between the bulb
and the foot of the tube - which most people thought was not possible - and realised it
together with the head of the glassblowing department, Haans, and developed it into a
method that could be used in the factory without the help of glassblowers . The choice of the
type of glass for both the image iconoscope and the Plumbicon was based on the
requirements that were imposed by these difficult constructions, in which tensions could not
be avoided, and on the optical requirement for a completely f1at screen."
Keeping the target material free from pollution was another problem.
These problems could only be solved by cooperation between researchers
and production workers .
The development of the Plumbicon shows the importance of a
piecemeal rationality: microtechnological designs like this one are the
result of intensive planning and coordination of R&D activities; each
design, however, can be considered as a new starting point for the
planning of activities that lead to further improvements in different fields:
in this case in electronics .
To gain more insight into the kind of coordination, required with
regard to organisation, steering and planning, we will deal with the
developments which led to a successful invention in the Philips Lab .
It would be wrong to look at these activities as a straightforward steering
towards a goal that was weIl defined and known from the outset.
The perception that a certain path will be a dead end sometimes creeps
up slowly after various analyses of the S-, T-, M- and P/J-factors.
In our historical-methodological study we did indeed come across such a
path, which turned out to be a dead end only after some time : the horne
cineac .
5. THE CONTEXT OF THE PLUMBICON INVENTION:

THE APPROACH OF THE PHILIPS LAB
5.a . The Home Cineac Program and the Inf/uences 0/ Msfactors

on Philips R&D
After 1936 Philips published hardly anything on pickup tubes. The only
person to publish articles in the 1930s was Rinia , who obtained many
patents for Philips. He was the man with the "golden fingers" : his designs
were excellent. When electronic pickup tubes had already been function-
ing for a long period (since the introduction of the iconoscope) he
developed a Nipkov disk that performed better than the best iconoscope
of that time." Apart from an article on this disk, nothing was published
by the Philips Research Lab in this area . Only in 1951, after a silence of
about 10 years, was a better version of the image iconoscope announced.
What was the reason for this gap in publications?
Here the myth of the both ingenious and strange Gilles Holst, who was
director of the Philips Lab between 1914 and 1946, seems to provide the
answer ." According to that myth, he seems to have had almost no confi -
dence in the development oftelevision. Some documents ofthe Philips Lab
archives seem to confirm this myth about Holst: he seems to have been such
a genius that he let people work on projects in which he did not believe
hirnself; individual persons would, more or less against their will, work on
television, whereas Holst hirnself seems to have been much more in favor of
the idea of ahorne cineac. This horne cineac would, every day, deliver the
news to subscribers by way of a film that they could watch whenever they
wanted. In this way people could escape from the tyranny of a dominant
television in their hornes. Fear of this tyranny would fit in with the per-
sonal convictions ofHolst, who held explicit ideas on humane technologies.
However attractive this image of Holst's ideas may be for explaining
television research at Philips, we think that other explanations are more
realistic. Holst did, in fact, very carefully watch the developments with
respect to the S- and T-factors, which he combined with the M- and P/ J-

factors we have mentioned.
The Philips archives show that a large number of people with a high
authority within the firm concentrated their research work on television .
Besides, it turns out that Holst brought together various disciplines in this
research: four electrotechnicians, two chemists and one physicist. It is
highly improbable that such a group would have been formed against
Holst's will." A crucial position was held by Rinia, who constantly
produced concrete designs and thereby kept Holst informed on the
progress of technological know-how, On the basis of this information
Holst was of the opinion, until 1940, that there was only a TV market for
hobbyists. Since Philips always aimed at mass production, this was not an
interesting field. In this case we see how a design functioned as a link
between scientific, technological and market factors: Rinia's examples
"showed" what kind of a market corresponded with the given S- and T-
factors.
A central focus of attention were transmitter problems." TV signals
are not reflected by the ionosphere and therefore have a limited range.
This implies that television broadcasting needs more transmitters than
radio and consequently would be more expensive. Holst posed this
problem to some of his very experienced people. Here again he seems to
show his awareness of M-, S-, and T-factors. Another M-factor was the
price of a TV set, that exceeded the monthly income of an average
working man at that time. " This aspect might, however, be eliminated
through changes in the S- and T-factors that would permit mass
production. Meanwhile, in 1937, calculations showed that a (horne)
cineac set would be less expensive than a TV set. In the research lab there
was a group, having Dippel as pacemaker, that believed in the market for
the horne cineac. Hol st did not have adefinite opinion about this product,
either, and decided that a bett er combination of S-, T-, and M-factors
might in due course settle the matter. In a confidential report of June 1944
on the horne cineac the Philips management was asked to take steps "for
the development of devices belonging to [the horne cineac] and
preparations for commercial activities". In the appendix the report
referred to American publications from 1939and said that Philips offered
ahorne cineac which "compared to television fis] something different and
superior in many respects" .3 6
On the other hand, from 1938 onward various analyses had been
carried out on the influence of TV sets in hornes. Attention was also paid
120 ANDRIES SARLEMIJN AND MARe DE VRIES
to the specific requirements for TV programs, compared to radio

programs." In 1938 a brochure with a survey ofthe areas in which Philips
was active explicitly mentioned television technology. It also Iisted the
technologies that still remained to be developed and here the need was
emphasized to work on projection television, because the small 20.25 cm
sets would be too tiring to watch. M-factors were dealt with as weIl, for
the account ended with the words :
not only will the programs cost considerable amounts of money, but the limited range of the
transmitters will cause the need for a great number of transm itters. The expensive
broadcasting and the high purchasing costs will make it extremely difficult to build a
commercial exploitation. And yet ... television will come . It will just take a number of
years."
This was a realistic insight into the given factors of that time. In 1941 a
special committee within the Lab stated that television would become a
device that would bring together good music, horne amusement and the
newspaper. For the horne cineac, however, Philips would have to venture
onto a market with which it had no connection: the movie world .:" One
year later Philips started experimental television broadcasting in
Bindhoven.:"
Altogether these facts support the idea that Holst undertook the
necessary preparations for the research program which ultimately yielded
successes Iike the Plumbicon.
The fact that, in the early 1940s, no final decision was taken in the
horne cineac versus television issue seems to be sufficiently explained by
the lack of concrete perspectives on S-, T-, and M-factors. Those
responsible were unable to base any decision on certainty, so both projects
were kept going until further information became available.
5.b . Management Changes in the Philips Lab and the First Successes:
New Iconoscopes and the Use 01 Lead Oxide in Vidicons and
the Plumbicon (1962)
Philips researchers had never been involved in developing pickup tubes

like the orthicon. From the preceding analysis the reason is clear: because
of the P/J-factors the orthicon was regarded as a typically American
design. The iconoscope, however, was considered a more promising
design, although it too was American. Philips kept its people working on
the iconoscope to discover a better solution to its secondary emission

problem: this solution would have to comply with the European
standards. This strategy was changed after the invention of the vidicon in
1950, which seemed an even better starting point for further
developments; the vidicon was regarded as a more "ripe idea". 41 The new
strategy led to the invention of the Plumbicon in 1962. Both the change of
strategy and its success must also be seen in the light of managerial
reorganizations at an earlier stage.
Holst retired in 1946. He was succeeded by three directors: the
theoretical physicist Hendrik Casimir, the physical chemist Evert Verwey
and the engineer and inventor Herre Rinia. Rinia was made responsible
for television. In 1948he received a telegram from Bouwman announcing
that the American market seemed to be ready for television.? Now H.
Bruining was made group leader of the television group, to which
belonged, among others, P . Schagen and J .C. Francken. His successor as
group leader was Haantjes.
In 1951 P . Schagen presented a thesis on the image iconoscope at the
Amsterdam Free University.? His ideas were elaborated in other
publications." This Philips image iconoscope sterns from the tradition
that Rinia kept going: the improvement of the iconoscope .
At the same time, however, Schagen and others worked in a different
direction. In 1954 they published a paper on the first Philips
photoconducting pickup tube, for which specially prepared lead oxide
was used. The choice of this material was based on its property of easily
absorbing X-rays: the photocurrent is caused by X-ray quanta. The
principle that was used allowed one of the other researchers, L. Heyne, to
write a doctoral thesis on the photoconductive properties of lead-oxide
layers." The new tube would turn out to be the immediate predecessor of
the Plumbicon.
In 1954 E.F. de Haan was recruited (by Haantjes) and added to
Bruining's group. Also involved in the research were: J .R. Boerman,
J .H .J . Maartens, T.W. van Rijssel, P .P.M. Schampers, T.G. Schut and
in, 1959, A. van der Drift. Of this group Van der Drift, Oe Haan,
Schampers and Schut belonged to the Philips Lab; Boerman, Maartens
and Van Rijssel belonged to the "Electrical Tubes" Production Group.
Heijne's invention did not yet mean the end ofthe iconoscope tradition
in the Philips Lab. A new improvement was reached in 1955 with the
scenioscope of Schagen, Boerman, Maartens and Van Rijssel. In this
design the isolating mica plate of the iconoscope had been replaced by a
glass plate with a limited conduction. The charge was not emitted, but
leaked away through the target. The same idea had been used in the
vidicon and it reappeared in the Plumbicon.
This broad range of research activities around various types of
semiconductors hints at a managerial organisation that facilitated
concentration on the set of requirements to be formulated for a weIl
functioning tube.
The Plumbicon was revealed to the public in 1962. Although its
invention has always been ascribed to one of the Lab's staff members ,
E.F. de Haan, it was the acknowledged result of a weIl-planned, collective
effort. The amount of work involved and the organization of research in
the Lab rule out serendipity-like and unplanned discovery by one single
person . No doubt Oe Haan played a vital role in the R&O activities and
even more so in the technical applications of this tube, but in his
announcement he cited a 1954 article by Heijne and others on their work
on the pickup tube with a light sensitive layer of lead monoxide.
The advantage of lead monoxide is that it can be doped so that it
becomes a so-called PIN diode, consisting of three layers:
- a P-Iayer,
- an intrinsic semiconductor I-Iayer, and
- an N-Iayer.
The PIN-diode together with the tin layer can be seen as the resistancel
capacitance combination that we already saw in the vidicon. The electron
scanning beam has the function of charging the P-Iayer. The diode blocks
the current when no light hits the target. That way there can be no dark
current, which, of course, improves the quality of the image.
6 . CONCLUDING REMARKS
6.a. Piecemeal Rationality during R&D-processes:

From One Design to the Next
Oe Haan and his colleagues published the principles of the Plumbicon in

1964.4 6 There they listed the set of requirements which the new tube had
to fulfil:
The Plumbicon combines small size, simple construction and easy operation with a low dark
cur rent , high sensiti vity , high speed of respon se, and good resolution . The Iife of the
APPLICA TlON-ORIENTED RESEARCH 123
"Plumbicon" is no shorter than that of other studio-quality tubes . It also produces striking
advantages particularly in colour television and in X-ray television set-ups ,
As a kind of thought experiment we now pose the following question : did

Nipkov or Baird know this set 01 requirements'l Of course not. Baird
thought of a market for people who wanted to broadcast simple
geometrical figures. For Philips, according to Holst, that was not an
interesting market.
This somewhat "absurd" question is aimed at revealing the dilemma
between planned R&D activities and hope for coincidence. That on the
one hand the serendip ity thesis can not be adequate, is evident from the
previous sections: millions or billions of dollars are spent on an
innovation like the Plumbicon. Besides that, each design is built on a
predecessor, for which an "arrny" of different scientists and technologists
had been mobilized. Only when we abstract from those efforts and
investments can we can talk of a chance event, one that occurs to a lucky
person .
However , a rejection of the serendipity thesis does not mean that we
are assuming a te!eological rationality in the complete chain of designs.
Planning is concerned with intermediate steps. After each step an
inventory is taken of the present situation, with all its remaining
uncertainties and newly gained knowledge. The transition from one
design to the next requires a comparison of the insufficient S- and T-
factors with the possible M- and PIJ-factors. It is then possible to
reformulate the set of requirements. Baird, for instance, thought he
already saw a market for his design. It cannot have been a large market:
just that of some hobbyists - which Holst would not even call a market at
all. Rinia with his designs in the 1930s made clear to his superiors what
could be attained on the basis of the available S- and T-factors. It was up
to the management to decide when these would fit in with their concept of
mass production and mass markets.
In 1940Holst brought together enough know-how to develop a further
understanding of the essential problems. After Rinia took over
responsibility for television from Holst in 1946, and after Bouwman had
observed the market in the USA, Philips was convinced of the potentials
of the TV market in Europe. In the 1950s much fundamental material
research got under way and experience technological developments were
pursued which, at the same time, widened the range of activities and
helped to concentrate the effort.
Of course, the rhetorie of smoothly planned R&D - without dead ends,

technieal failures and financial disasters - often comes to the fore when
the management of successful industries explain their victories. However,
the common sense rationality of these accounts is based on a somewhat
myopie hindsight. When looking back over a longer time span, a different
pieture emerges. The development of pickup tubes confirms the idea that
only intermediate steps can be planned, and not the entire marehing route .
The changes in S-, T-, M-, and P/J-factors were evaluated after each
design. Depending on the expectations and prognoses , a new plan was
made and various kinds of know-how were mobilized, depending on the
perspectives. This is what we have called here a piecemeal rationality: the
development consists of separate pieces, or designs, whereby each design
realizes only limited goals. There is not one rationality that guides the
whole chain of designs.
6.b. Influence 0/ Nat ionalism on R&D
Wehave seen how nationalistic interests can steer R&D in a certain

direction : the discord on the number of lines to be chosen caused different
countries to develop different types of tubes . That does not, however,
stimulate mass production. In a market situation with sharp competition
one has to try to break down national barriers. That was the reason why
a number of countries came to an agreement , so that one type of tube (the
ieonoscope) was usable for all of them, and later on the nationalist
differences became irrelevant because one type of tube (the vidieon/
Plumbicon) could be used by both Anglo-Saxon and continental
countries.
6.c. Design Methodological Factors and Push & Pull Theories
It is hinted at, by Casimir, for instance, that especiaIly in the field of

mierotechnology, S-factors always have to precede T-factors, whieh in
their turn should stimulate M-factors to open new markets . This is the
weIl known technology push thesis, and it indeed is not entirely
unrealistie: the theory of materials (semiconductors) is a framework that
one has to take into account. But with the television pickup tubes S-
factors only played a vital role in one distinct phase and we should not
exaggerate their earlier influence. Braun's tube, the diode and the triode
were largely based on experience without an advanced theory of electrons.
These devices were sufficient for an empirical development of the pickup
tube. Only with the finishing touch, in the vidicon and the Plumbicon, did
solid state and surface physics come to play an important role in the
search for suitable materials.
According to oral tradition in the Philips Lab, Holst would not have
had a clear view of market developments. However, when we reconstruct
the technological possibilities and problems of the 1930s and the
management philosophy of the Philips corporation of that time, it
appears that his hesitation was justified: the S- and T-factors were
inadequate to meet the markets on which Philips wanted to operate.
The dilemma between push and pull, is at least in our case, a pseudo-
problem. The two concepts in no way relate to the reality of innovative
activity, and their simple dichotomy tends to obscure our understanding
of innovation. The ever changing relations between So, T-, M- and PI J
factors deserve our full attention when analyzing the "rationality" of
design.
Eindhoven University 0/ Technology
NOTES
* The authors want to thank Dr L. Heijne and Dr T.S . te Velde (both have worked in the
Philips Research Laboratory on crucial projects of television technology) for their
comments on earlier versions of this article, and the employees of the Philips Concern
Archive for their cooperation in identifying resources for this study .
l Abbreviated to "Philips Lab".
1 Towler [1986], p. 67
J Efforts to raise the analysis of normalization processes to a specific science (the so-called
"Normierungskunde") did not succeed so far ; for the formulation of such an effort, see
Berg [1974].
4 We find this vision for instance in all publications by P . van Andel (one ofthem is [1989]).
, This opinion dates from the Sixties, when revolutionary students and their ideologists
thought that corporations would pursue their own interests and not society's interests. A
recent and more explicit version of this idea is found in the publications of the co-called
Starnberger School with as its most important representatives : Böhme, Van der Daele and
Krohn; for the publications the reader can consult the list of references. During a conference
in 1986 in Erlangen , Germany, a lot of "acadernics" appeared to hold these ideas (see
Sarlemijn, [1987]).
• The programrne, in which this concept fits, has been expla ined in Sarlernijn, (1990).
7 For more details of this tripartition , see Sarlemijn, (1984) and (1990).
8 After 1895-1905 the equations of classical physical theories became calculation methods
for the engineering sciences : the discussions about theoretical concepts (like the aether) had
by then been decided. This distinction between physical theories and engineering methods is
explained in the works of Casimir (1958), (1979) and [1983J. See also note 42.
9 Nipkov's German patent received number 30105 in 1884. Nipkov's method has been
explained extens ively by Zworykin and Morton [1954J and Csorda [1985J.
10 Figures land 2 have been taken from Holm [1956J. The recording of the two women is
probably Rinia's work, who experimented with the Nipkov disk (see Rin ia et al. [l937J,
[1938J and (1939) and also note 20).
11 See Na/ure 78 [1908J, p . 151, Journal of the Röntgen Society 8 [1912J, pp . 1-15 and
Wireless World 14 (1924), pp . 51-56, pp . 82-84, pp. 114-118. See also the judgement on
Allan A . Cambell Swinton's merits in Zworykin and Morton (1954), pp . 245-247 . In'
Schagen et al. [l951J, p. 71, Cambell's designs are mentioned as "strange drawing" .
12 For areport of Baird's experiments, see Nature (1925) 115, p. 505. Baird's contribution
can be found in Tiltman [l974J .

13 See Jenkins [1925J and a description of his system in Wireless World 18 [1926J, pp .
642-645 .
14 This expression, which we find in Druyvesteyn's letter (PAD , (1967», has become a weil
known saying to indicate Holst's and Ca simir's management practice in the Philips Research
Lab .
15 Abramson [1974J, p. 7.
16 The demonstration according to Druyvesteyn (PAD, [1967]), has been discussed in
Radio -Expresse of Dec. 1928.

17 Zworykin [1934J, Zworykin and Morton [1954J.
18 Van der Mark (1936), p. 325.
19 See Towler (1986).
20 Berthillier (1947), Zworykin and Morton (1954).
21 Kerkhof and Werner (1951).
22 lams and Ro se (1937).
23 Kerkhof and Wern er (1951), p. 61.
24 This conviction is expressed in Schagen, Bruining and Francken [1951J, p. 73) in the
following way: "According to these possibilities, pickup tubes are divided as folIows : (I)
pickup tubes with low electron velocities ("low velocity tubes"), in which the target is
stabilized on the cathode potential (that is why they are also called " cathode stabilized
tubes" , abbrev. CPS tubes), and (2) pickup tub es with high electron velocity ("high velocity
tubes") , in which the target is stabilized on the potential of the collector (e.g . 1000 V). To
the first group belongs the image orthicon which is mainly used in the USA, to the second
belongs the image iconoscope, which is preferred in Europe . One of the reasons for this
preference has to do with the great number of fines for which was decided on the Western
European continent (625, in France 819) . With high electron velocity one can better fulfil the
high requirements the definition of such a great number of lines poses to the focusing of the
scanning beam" (italics Sarlemijn and De Vries) . An earlier internal report (PCA (1948»,
probably written by Lewin , contained another vision: the 567 lines and 25 irnages per second
system is not technically different from the American system, but implies an economical
optirnum, because increasing the number of Iines would increase costs without appreciable
improvement of the image qual ity, and reducing the number of Iines would not save much
cost, but would seriously worsen the image quality , Again we see the problem of the number
of transmitters that Holst put forward time and again .
as see Albers [1953].
2. Although from 1948 onward all doubt in the Philips Lab disappeared (see Sarlemijn
[1990]), yet for a long time in the Fifties one was not confident about the speed of growth
and the final size of the market. This largely depended on the extent to which programmes
would become bette r. In The Netherlands the production of programme s was rather
primitive at that time (see Alders [1953]).
21 See Abramson [1974]. p. 7.
2. Plumbicon is the trademark for a certain type of vidicon. That is the reason why the name
- according to the intention of the patent application - should be written with a capital P .
2. Philips Archives. File Plantinga NL 661, first report.
JO L. Heijne in a letter to the authors, 14.01.199J.
JI This judgernent can be derived from the following publication s: Rinia [1938], Rinia and
Dorsman [1937]. Rinia and Leblanc [1939] . It has been confirmed in several interviews.
J 2 This myth forming has something paradoxical about it: it usually sterns from deep
adrniration, but not infrequently hurts the authority concerned .

JJ lvo Blanken of the Philips Archive pointed out that Holst. though with sepsis, caused
resentment during management meetings on the television. That does not clash with our
thesis, which has three aspects: because of the So, T-. M- and P/J-factors Holst estimated
television to be not ripe for Philips' aspiration for the market. He kept watehing those
factors accurately without being led by speculative thoughts . He worked on the mobilization
of forces as soon as he believed a realization would become possible, putt ing the most
competent workers on those matters (e.g. the problem of the number of transm itters) from
which he expected the most problems.
J4 This is confirmed by nearly all PAD s that have been included in the list of references.
J 5 According 10 a rnessage of Philips Lamper (Oslo), dated September 19. 1929 (PAD), a
television set would cost about 700 Marks . Calculat ions, made in Eindhoven , resulted in an
amount of Dfl. 1.000.-.
J' PAD (1944) .
J 1 PAD [1938] .
J. Philips' Bedrijven .... etc.• [1938]. p. 73.

J' PAD [1941] .
40 An experimental transmitter is mentioned in Van der Mark [1936] . PAD [1948]says: "To
gain the necessary experience as soon as possible, the Philips Research Laboratory put into
operation a 20 kilowatt television transmitter, that willtransmit programs and films for 4.5
hour s a week. In Eindhoven and surroundings a Iimited number of receivers will be
mounted, of a simple type with cathode tubes for direct reception as weil as projection -
receivers that give an image of about50 x 50 cm" . In PAD [1951]we read : "Already before
1939 by means of a moving caravan Philips gave television demon strations in most
European capital s. The results thus gained were very good and added to Philips ' prestige in
the area of television" .
41 We meet this concept of a "ripe idea" also in a theoretical context, namely in the works
of H .B.G. Casimir; he discusses theories that are ripe for fundamental developments in an
industriallab. In (1979) Casimir answered the question when quantum theor y was "ripe"
enough to serve as a starting point for industrial research as folIows: when the basic
equations had been found . Before that they had to be the subject of academic research.
Indeed in about 1930 in the Bell Lab as weil as in the Philips Lab quantum theory and
quantum mechanics became the object of intensive studies.
'2 This document is reproduced in Sarlemijn [1990).
'3 Schagen [1951].
•• Tijdschrift van het Nederlandse Radiogenootschap 16 [1951], p . 236 and Schagen,

Bruining and Francken [1951].
.. Heijne [1960].
' 6 De Haan, Van der Drift, and Schampers [1963/64] .
REFERENCES
Abbreviation: PTR = Philips Technical Rev iew
Abramson , A.: 1974, Electronic Motion Pictures, New York, etc.: Arno .
Abramson, A.: 1987, The History of Television, 1880 to 1941, Jefferson, etc.: McFarland .
Alders, J .c.: 1953, Opmars der electronen, Assen: Born.
Andel, P . van and Andreae, B.: 1989, Serendipiteit, de paradox van de ongezochte vondst,
Groningen: RuG.
Berg, S.: 1974, Normung, Forschung, Entwicklung (Normungskunde, Heft I, hrsg. v.
Deutschen Normenausschu sz). Berlin, usw.: Beuth.
Berthillier, M.: 1947, 'The Eriscope', Le Vide 2, pp. 355-359.
Böhme, G., van der Daele, W. and Krohn, W.: 1973, 'Die Finalisierung der Wissenschaft',
Zeitschrift für Soz iologie 2, pp. 128-144.
Böhme, G., van der Daele, W. and Krohn , W.: 1978, ' Finalisierung Revisited' , in Böhme,
G. (Hrsg), Die gesellschaftliche Orientierung des wissenschaftlichen Fortschriftts,
Frankfurt alM : Surkamp, pp. 195-250.
Burns, R.W.: 1986, British Television: The Formative Years, London : Peregrinus .
Campbell Swinton, A.A.: 1908, ' Distant Electron Vision', Nature 78, p. 151.
Casimir, H.B.G.: 1958, ' Certainty in the Exact Sciences', in Verslag van de plechtige viering
van het 150 j arig bestaan der KNA W . Amsterdam : Noordhollandsche, pp. 243-251 .
Casimir, H.B .G.: 1979, De kringloop van natuurkunde en techniek in de 20e eeuw,
Haarlem: Holl. Mij der Wetensch.
Casimir, H.B.G.: 1983, Haphazard Reality, Half a Century of Science, New York, etc.:
Harper and Row.
Csorda , I.P .: 1985, Image Tubes, Indianapolis : Sams.
De Haan, E.F .: 1962, 'The Plumbicon , a New Television Camera Tube' , PTR 24, pp.
57-58 .
De Haan, E.F ., van der Drift, A., and Schampers , P.P .M.: 1963/4, '''The Plumbicon", a
New Television Camera Tube' , PTR 25, pp. 133-161.
Doorn , A.G . van: 1966, '''The Plumbicon" Compared with Other Television Camera
Tubes', PTR 27, pp. 1-14.
APPLICATION -ORIENTED RESEARCH 129
Everson, G.: 1974, The Story 0/ Television: The Life 0/ Philo T. Farnsworth , New York:
Arno .
Farnsworth, P .T.: 1934, 'Television by Electron Image Scanning', J. Frankfin Inst. 218
(October), pp . 411-444.
Fink, D.G .: 1960, Television Engineering, New York: McGraw-HilI.
Francken , J .C. and Bruining, H.: 1953, 'New Developments in the Image Iconoscope'. PTR
14, pp . 327-335.
Heijne, L. 1960: Photoconductive Properties 0/ Lead-Oxide Layers (Thesis) Amsterdam :
Free University.
Heijne, L. : 1991, Letter to the authors in reaction to a draft version of this articie. (14
January).
Heijne, L., Schagen, P ., and Bruining, H.: 1954, ' An Experimental Pickup Tube for
Television with Application of Photoconduction' , PTR 16, pp. 43-45.
Holm , W.A.: 1956, Televisietechniek, Amsterdam : Meulenhoff.
larns, H. and Rose, A.: 1937, 'Television Pickup Tubes' , Proc. IRE 25(August), pp .
1048-1070.
Inst , of Electronic Engineers: 1986, International Conference on the History 0/ Television
- From Early Days to the Present. London .
Jenkins, C.F .: 1925, ' Vision by Radio Photographs', Rad. Phot. Nat . Cap. Press.
Kerkhoff, F. and Werner , W.: 1951, Televisie, inleiding tot de physische en technische
grondslagen van de televisietechniek met een uitgebreide beschrijving van
schakelschema 's, Amsterdam : Meulenhoff.
Mayo, B.J .: 1986, 'Early Development of Electron Guns and Cathode Ray Tubes for
Television', in International Conference on the History 0/ Television - From Early Days
to the Present. London: Inst. of Electron ical Engineers, pp . 69-72.
McIlwain, K.: 1939, 'Survey of Television Pickup Devices', J. Appl. Phys. 10(7), pp.
432-442.
(PAD) Philips Archive Documents:
- 1929, Message from the Norsk Akt ieselskap 'Philips Lamper' , Oslo, 19 September .
- 1937, 'Televisie express, de toekomst der televisie', ir. H.H . Heerema compares
advantages and disadvantages of the horne cinema and the television and also makes
price comparisons (PCA 811.3).
- 1938, ' Waarom geen televisie in Nederland ', analyses by ir. M. Leeuwin (PCA 811.3 +
814.3).
- 1939, Van der Pol's deliberations with respectto possible places for transmitters for TV
in the Netherland s, message to dr. A.F. Philips (PCA 911.3 + 814.3) .
- 1940 (28.8), Division radio research, message to those who work on television in the
Research Lab (PCA NL 608).
- 1940, Annual survey of the FM research in Eindhoven with appendices on the progress
in 1941 and 1942, by B. van der Pol (PCA 811.2) .
- 1941 (30.10), 'Celophane', report of the commission Department (NL 828).
- 1944 (30.6), Advice of J .F. Berger and ir. J .P . Bourdrez concerning the horne cineac.
- 1946, 'Possibilities for television', text of a radio broadcast by ir. M. Leeuwin on 14
August 1946.
- 1948 (8.1), 'Television in Europe', ir. M. Leeuwin compare s the advantages and
disadvantages of various line systems.
- 1948, 'Sorne ideas with repect to a Benelux coperation in the area of television' , by
Leeuwin (PCA 811.3 + 814.3).
- 1951, 'Philips and television ', Philips Press Service.
- 1967(12.12), letter of Prof. Dr. M.J . Druyvesteyn to Van Lokeren Campagne Public
Relations, N.V. Philips.
- Interview (tape 2) with dr. Haantjes by dr. Bruining and ir. StieItjes.
Philips' Oloeilampenfabriek: 1938, Philips Bedrijven, een korte beschrijving van .. . .
Eindho ven.
Puetz , J. : 1978, Television. S.l. : Verlaggesellschaft Schulfernsehen.
Rinia, H .: 1938, 'Television with Disc According to Nipkov and Interlinear Scanning', PTR
3, pp. 289-295.
Rinia, H. and Dorsrnan, C: 1937, 'Television Installation with Disc According to Nipkov",
PTR 2, pp. 72-76.
Rinia, H. and Leblanc, X.: 1939, ' Nipkov's Disc', PTR 4, pp. 46-52 .
Rose, A. and larns, H. : 1939, 'The Orthicon , a Television Pickup Tube ' , RCA Rev, 4, pp.
186-199.
Sarlernijn, A.: 1984, Histor isch gegroeide relaties tussen natuurwetenschap en techniek;
Eindhoven : TWIM .
Sarlernijn, A.: 1987: 'Analogy Analysis and Transistor Research' , Methodology and Sei.
20(3), pp. 40-61
Sarlernijn, A.: 1987, ' Konvergenz im Bezug auf Planung der Forschung' , in Burrichter, C.
and Lauterbach, O. (Hrsg), Wissenschaftsforschung im internationalen Vergleich,
Erlangen : DOF, pp . 227-281.
Sarlernijn, A. : 1990, Het ontwerp, spil van de techniekkultuur (inaugural lecture).
Eindhoven : University of Technology .
Sarnof', D.: 1939, 'Probable Influences ofTeievision on Society', J. Appl. Phys. 10(7), pp ,
426-431.
Schagen, P .: 1951 , On the Image Iconoscope, A Television Camera Tube (thesis).
Amsterdam : Free University.
Schagen, P ., Bruining, H., and Francken, J .C .: 1951 , 'The Image lconoscope, a Pickup
Tube for Television', PTR 13, pp. 71-85 .
Schagen, P :, Boerrnan, J.R ., Maartens, J.H.J ., and van Rijssel, T.: 1955, 'The
"Scenioscope", a New Pickup Tube for Television', PTR 17, pp. 173-182.
Smith, J .M. and Stewart, J.: 1986, 'From Vacuum Tubes to Solid State Imagers - the Past,
the Present, the Future', in International Conference on the History of Television - From
Early Days to the Present. London : Inst. of Electronica1 Engineers, pp. 73-76 .
Stern , R.H. : 1950, The Federal Communications Commission and Television; The
Regulatory Process in an Environment of Rapid Technical Innovation (thesis).
Cambr idge, Mass.: Harvard University
Teer, K.: 1986, ' Looking Back at Distant Vision: Television Technology from 1936-1986',
PTR 42, pp. 297-311.
Thiele, R. and Fix, H.: 1959, ' Ein vergleichende Betrachtung der heute verfügbaren
Fernseh-Bildaufnamerohren', Radio-Mentor 25, pp. 448-452.
Tiltman , R.F .: 1974, Baird of Television : The Life Story of John Logie Baird. London :
Seeley.
Towler, 0 .0. : 1986, 'Milestones in Television Pickup Performance', in: International
Conferenceon the History of Television - From EarlyDays to the Present. London : Inst.
of Electronical Engineers, pp. 64-69.
Van der Daele, W., Krohn, W., and Weingartner, P. (Hrsg.): 1979, Geplante Forschung.
Frankfurt alM : Surkamp.
Van der Mark, J.: 1936, 'An Experimental Television Transmitter and Receiver', PTR I,
pp. 16-22.
Van der Mark, J. : 1936a, 'Television' , PTR I , pp. 325-330.
Weimer, P.K., Forgue, S.V., and Goodrich, R.R .: 1950, 'The Vidicon Photoconductive
Television Pickup Tube' , Electronics 23, pp. 70-73 .
Zworykin, V.K.: 1935, 'The Iconoscope' , Proc. IRE 22(January), pp. 16-32.
Zworykin, V.K. and Morton, G.A.: 1954, Television, the Electronics of Image
Transmission in Colorand Monochrome. New York: Wiley, London : Chapman and Hall
(2nd ed.).
ROBERT FOX AND ANNA GUAGNINI
LIFE IN THE SLOW LANE: RESEARCH AND

ELECTRICAL ENGINEERING IN BRITAIN, FRANCE,
AND ITALY, CA. 1900
This paper is a reflexion on an assumption that has run through studies of

industrial technology ever since the First World War . The assumption,
broadly stated, is that in the age of science-based industry a strong and
preferably expensive commitment to research and development is an
essential ingredient of a nation's technological prowess . The corollary is
that technological success and, by a deceptively easy extension, economic
success have come to be seen as being dependent on a capacity for
autonomous innovation. A serious commitment to research has assumed,
in the process, an almost symbolic and, as we shall argue, exaggerated
importance. Among historians of the period that concerns us here - the
late nineteenth and early twentieth centuries - it has come to be regarded
as one of the main touchstones dividing the countries we traditionally see
as the technological pacemakers - in particular Germany, the U.S.A., and
Switzerland - from those, of which Britain, France, and Italy are typical,
which are usually portrayed as limping along in their wake.
The kind of analysis that concerns us is evident in many case studies of
the relations between science and industrial performance. John Beer's still
unrivalled account of the rise of German dyestuffs manufacture is typical
in insisting that one of Germany's cruciaI advantages Iay in the
availability of chemists trained in research and in the engagement of those
chemists either as consultants or in in-house laboratories. I Likewise, the
more general discussions of economic historians, from David Landes's
Prometheus Unbound to the recent work of David Mowery, have tended
to give Iaboratory-based scientific research a central role in their
explanations of the development of industrial technology in the twentieth
century.' The obverse of the coin is, of course, that where research has
been absent, innovation is supposed to have flagged, heralding a natural,
even inevitable decline. The point is very clear in Mowery's statement
about Britain . For Mowery, a low level of industrial research was among
the main causes of Britain's poor innovative record and hence of her
weakening economic performance since 1900.3
133
P. Kroes and M . Bakker (eds.), Technological Development and Science in the lndustriol
Age, 133-153.
© 1992 Kluwer Acodemic Publishers.
134 ROBERT FOX AND ANNA GUAGNINI
THE FAST LANE AND THE SLOW LANE IN ELECTRICAL TECHNOLOGY
In turning to the particu1ar sector we have chosen for this paper - that of
electrical manufacturing and supply - we do not wish to challenge the
main statistical data on which current assessments of economic success
and failure are based. The implications of the figures given by such
authors as I.C .R . Byatt and Peter Hertner, which demonstrate the
capacity of German manufacturers to dominate markets outside as weil as
within the Empire, are clear." It is beyond question that weil before the
First World War, possibly even by 1890, Germany, along with the U.S.A.
and Switzerland, was firmly entrenched in what we term the fast lane of
electrical technology, while Britain and France drifted into a distinctly
slower lane from which escape became increasingly difficult as the gulf
between the leaders and the followers in the sector widened, as it did
inexorably, especially between 1900 and 1914. This slow lane was one in
which we also find Italy. For, despite the remarkable advance of the
economy of Italy from about 1890, technological dependence remained a
dominant characteristic of Italian industrial development until after the
First World War.
Our scepticism, then, concerns not the data themselves but rat her the
appeal to research as an explanation for them. In our view, the tenets
underlying such an appeal demand a conscious scrutiny which they have
all too seldom received. How justified is the assumption that improved
economic performance can only be achieved through a major engagement
in independent innovative activity? And is it invariably the case that
innovation of this kind emerges from the paradigmatic royal road of
research pursued, in the conventional late-twentieth-century manner, by
laboratory-based scientists?
Our answers to these questions are set in a very specific context. Our
concern is primarily with Western Europe and exclusively with the period
before the First World War, when electrical manufacturing and
installation emerged as a classic instance of a science-based industry. It is
symptomatic of the difficulty of formulating answers that even in the
seemingly innocent phrase "science-based industry" there lie teasing
historiographical difficulties, especially with regard to the period we are
discussing. For it is assumed, all too easily in our view, that the
description "science-based" is necessarily synonymous with "research-
based" and with activities geared to the quest for novelty pursued in the
laboratory. That assumption, which seems to draw such plausibility as it
RESEARCH AND ELECTRICAL ENGINEERING 135
has from an extrapolation backwards from the period after 1918, is

perilously attractive both for historians of science and technology and for
economic and business historians. But it is one that we cannot accept.
It is easy to see why the assumption is so easily made . For ever since the
1890s contemporaries both in the academic world and in manufacturing,
and in the countries of both the fast lane and the slow lane, have displayed
an almost breathless admiration for what they have chosen to portray as
a gradual rapprochement of science and industry. The official catalogue
for the German electrical section at the Universal Exhibition in Paris in
1900 ascribed Germany's leading position to, above all else, the presence
of almost two thousand scientifically trained graduates of the Technische
Hochschulen in the main German companies.' Similarly, in the slow lane,
the view that research and education were the foundations of Germany's
(and, later, the U.S.A.'s) industrial success appeared repeatedly in the
public statements of scientists and educationists. In Italy, the point was
forcibly made by the chemists who gathered in Turin for the inaugural
meeting of the Associazione di Chimica Industriale in 1902.6 At about the
same time, in Britain, an unprecedentedly clamorous lobby in favour of
research began to make its mark in the debate on the reorganization and
expansion of higher technological education, notably (though by no
means exclusively) in the discussions that culminated in the launehing of
Imperial College." In France, too, vociferous champions of research,
most of them speaking from within the academic community, fostered a
mixture of genuine admiration and anxiety, as they harped on the
awesome superiority of the provision for laboratories in the German
Empire and elsewhere."
But such accounts have to be read with extreme caution. In particular,
it cannot be assumed that the conceptions of the role of science in industry
and of the functions of applied research were those which circulate in
debate today, The fact that only five per cent of the graduates of the
Technische Hochschulen working in the German electrical industry ab out
1900 were engaged in laboratories is one indication of the danger of
assuming that science has always made its most natural contribution to
industry in the pursuit of innovative research ." An even more telling
illustration is the campaign for the cause of industrial science that was led,
from the late 1890s, by the metallurgist and engineer in the French Corps
des Mines, Henry Le Chatelier. In arguing for the need for French
industry to open its arms to science and reject the unchanging rule-of-
thumb traditions of the shop floor, Le Chatelier was not primarily
concerned with the growth of research for innovation. He wished rat her
to see industry manned by a new breed of scientifically trained engineers,
and the national and industrial laboratories for whose establishment he
tirelessly campaigned were conceived first and foremost as settings for
quality control and testing and only incidentally as sources of
fundamentally new products or processes. 10
In Italy, a stance similar to Le Chatelier's was taken by Giuseppe
Colombo . Speaking from his powerful position at the head of the Istituto
Tecnico Superiore of Milan from 1897 to 1921, Colombo advanced a
programme for the modernization of Italian industry that drew on his
experience as a successful consuIting engineer, a member of parliament,
and the founder and first president of the Societä Italiana Edison. It was
a programme that gave pride of place to the introduction and diffusion of
existing technologies, rather than to the development of new ones .!'
Above all else, it required men who were able to assess both the
technological and economic merits of existing practices abroad and to
adapt those practices to the conditions prevailing in Italy. For this
purpose, engineers versed in the latest technologies were indispensable,
and under Colombo the Istituto Tecnico Superiore duly provided an array
of theoretical and practical courses oriented to the purveying of received
knowledge. The low priority that was accorded, by contrast, to the
fostering of fundamentally new departures was reflected in the virtual
absence of facilities for research."
It is important to reflect on the reasons why Le Chatelier and Colombo
chose to give innovation this secondary role and to conceive laboratories
and training in ways that may seem, in retrospect, unnecessarily
restricting. Were they simply being myopic? Were they failing to perceive
the key to industrial success that lay in innovative research and thereby
contributing to the perpetuation of the virtual imprisonment of their
countries in the slow lane? Both suggestions seem to us unconvincing. One
reason for this is that, although we readily admit that in electrical
engineering an indifference towards laboratory-based innovative research
undoubtedly existed in both France and Italy, we do not believe that
attitudes in countries in the fast lane were consistently different.
Certainly, the commitment to such research in, say, Germany or the
U.S.A. does not seem to have been strong enough for us to see it as a main
source of those countries' technological superiority. The roots of that
superiority lie in a far more varied mosaic of economic, political, and
technological factors which the literat ure on the history of industrial
technology, in particular work inspired by the insights of Nathan

Rosenberg, is only now beginning to uncover. 13
LABORATORY PRACTICE AND INDUSTRIAL SUCCESS
The scepticism we have voiced has overtones that will be immediately

familiar. In obvious ways, it draws part (though only part) of its strength
from the exploding revisionist literature on laboratory practice that has
circulated in recent years . One very pertinent conclusion that emerges
from Leonard Reich's history of research at the American company
General Electric, for example, is that the term "research" is itself
obfuscating.!" Reich describes a company laboratory at the beginning of
this century that was far more concerned with the protection of patents
and with quality control than with the quest for novelty. Despite the
determination ofthe company's Vice President, Edwin Rice, that General
Electric should maintain "a real scientific laboratory", the reality was
distinctIy different. The laboratory justified its existence by standard-
izing, enhancing reliability, and examining external patents that might be
purchased with a view to creating a kind of protective belt from which
rivals would be virtually excluded . And it was only late in the day, weIl
after the turn of the century, that General Electric successfully launched
a research laboratory geared to innovation. It was in this laboratory, of
course, that the important but lang and very expensive research leading to
the new metal filament for the electric light bulb eventually bore fruit. But
that success should not obscure the lateness of the emergence of this type
of research . Nor should it obscure the fact that the new, innovative thrust
in research would not have been possible but for the contribution to the
security and growth of General Electric that had been made by earlier
laboratory work of a far more routine, non-innovative kind .
Our scepticism has also been fed by a parallel revisionism directed at
academic rat her than industriallaboratories. David Cahan's book on the
Physikalisch-Technische Reichsanstalt, the biography of Lord Kelvin by
Norton Wise and Crosbie Smith, and Graeme Gooday's work on physics
laboratories in Victorian Britain all point to an emphasis on painstaking,
precise measurement that characterized both research and teaching in
physics in the later nineteenth century. IS This new literature has had the
effect of bringing such long-disregarded figures as Victor Regnault in
France and Friedrich Kohlrausch in Germany back into a perfectly
respectable main stream of nineteenth-century physics in which novelty

was not the only worthy goal. 16 It has also shown how much care
physicists, even of the stature of Heimholtz and Lord Kelvin, devoted to
precision measurement and standardization and how important that work
was for their contributions to the industrial applications of science.
As we have insisted, however, our unease about the exaggerated
importance that can so easily be attached to laboratory-based innovative
research as a mainspring of technological success sterns only in part from
these historiographical trends. For however profoundly the revisionism
has adjusted our perception of the character of the work that was
conducted in industrial and academic laboratories, it has not displaced
those laboratories from the central position they still hold in the histories
of successful companies. The existence of structured laboratory facilities,
substantial expenditure, team-work (as opposed to the efforts of
individual inventors), and a close, formalized interaction between the
laboratory and the factory all remain the classic ingredients of the
achieving firm .
In voicing this reservation, we should not wish to deny innovative
laboratory-based research the successes that it unquestionably had before
the First World War. Our concern is simply that a few obvious, though in
our view untypical, instances of major new departures, such as the metal
filament lamp at General Electric, should not be allowed to obscure the
less spectacular and often quite routine character of most of what was
done in the overwhelming majority of company laboratories, in the fast as
weil as in the slow lane. Here, the case of AEG, a large and consistently
profitable firm that made its way on the basis of only a modest and by no
means systematic commitment to fundamental innovation generated "in
house", is an outstanding example that contrasts sharply with the far
greater reliance on independent research in Siemens & Halske and even,
on a smaller scale, Schuckert ." It is typical of AEG that it preferred to pay
Walther Nernst a million dollars for the patent rights to the highly
innovative Nernst lamp, while concentrating the activities of its own
laboratories on a conceptually less exciting (though commercially very
profitable) programme aimed primarily at enhancing reliability, reducing
production costs, and adjusting products ever more precisely to the needs
of customers.
There is another, quite different reason why the "ideal type" of the
successful company with a heavy investment in innovative laboratory
research is misleading. This is that it diverts attention from the innovation
that could and did take place in the workshop or in a factory or generating
station. Here, the innovation was typically of a far less spectacular kind
and of an adaptive rather than a fundamentally novel character. Hut it
provided another channel by which science could enter industrial practice
and often produce more immediately profitable results . Indeed, it is
tempting to suggest that the key to the technological success of countries
in the fast lane of the electrical industry may have less to do with the very
visible world of the laboratories than with the far obscurer mechanisms
for the transfer of scientific knowledge through experimentation
conducted on the shop floor or at an installation.
The force of this suggestion lies in the location of the cutting edge of
electrical technology in the quarter of a century or so before the First
World War. The location was unequivocally in the exploitation of high-
voltage, polyphase alternating current. It was a world of large generators,
long-distance transmission lines, electric traction, and the electrification
of industry. In all of these areas, there were major new departures, but
they were seldom ones that originated in the laboratory, at least in the
laboratory as it is traditionally conceived . The laboratory that mattered
for the pioneers of high-voltage A.C. before 1914 was the workshop or the
site . Here, the activity was scientific to the extent that it involved
experiment, theoretical insights, sophisticated mathematical techniques,
and the systematic analysis of production methods. The workers in this
"laboratory" , however, were engineers rather than reconverted
physicists. They constituted a very different breed who combined
scientific competence with experimental interests focused firmlyon the
machines and installations themselves . They were the products of courses
oriented explicitly to electrical technology, albeit to an electrical
technology that demanded an increasingly refined grasp of physical
theory.
The contributions made by these men were at times highly innovative,
as the cases of Marcel Deprez, Nikola Tesla, or Michael Dolivo-
Dobrowolsky illustrate. Hut, more frequently, they were of a less
spectacular kind, affecting quality control, reliability, or production
costs, or the harnessing of electric power in new industrial contexts. The
case, recently described by Ulrich Wengenroth, of the exceedingly
intricate tasks that were involved in adjusting polyphase A .C . motors to
the complex requirements of a textile mill is typical." As Wengenroth
shows, the gearing and coupling mechanisms that were required in order
to ensure a uniform speed irrespective of the number of machines in use
were not fundamentally novel. But the quest for a solution called for
precise measurements of efficiency and a degree of mechanical ingenuity
and flair that it is all too easy for the historian to undervalue by
comparison with the purely electrical side of the task.
BROADENING THE CONTEXT OF INDUSTRIAL SCIENCE
The argument we want to develop from this discussion is that, in the fast
lane as much as in the slow lane, the contribution of science to industrial
practice cannot be fully understood without a c1ear appreciation of
changes of the kind we have described, some with their focus on such
tasks as trouble shooting and the adaptation of electrical technology to
customers' requirements, others yielding innovations in contexts other
than that of the conventional laboratory. If the argument is accepted, it
has significant implications for the continuing debate on the extent , and
the implications for industrial performance, of the technological gap that
opened between, say, Germany and the three slow-Iane countries -
Britain, France, and Italy - with which our own work has been chiefly
concerned. Britain , France, and Italy, after all, expanded their provision
for technical education to levels which ensured that about the turn of the
century the supply of manpower at least matched demand. In our view,
therefore, the basic technical knowledge and skills were not wanting, and,
on this score at least, we see no impediments to the development of a
technological strategy, founded on the assimilation and adaptation of
borrowed technologies, that offered every prospect of success.
The questions that emerge from this interpretation are of a somewhat
different kind from those conventionally asked in discussions of the
relations between scientific knowledge and industrial practice. They turn
less on the capacities of different countries to generate either an elaborate
provision for advanced technical education or independent programmes
of research and innovation than on their being in a position to exploit the
manpower and the facilities (including laboratories for such routine tasks
as testing and standardization) that were more generally available than
most of the current secondary literature would suggest.
Such questions are riddled with snares and deceptively cogent answers.
Quite apart from the temptation to seize on a low level of laboratory-
based innovative research as a satisfactory explanation of flagging
industrial competitiveness, it is all too easy to associate any signs of a lack
of interest in the more daring kinds of technologieal novelty with the

failure of a vaguely conceived entrepreneurial spirit. These answers, in
our view, will not do, for the simple reason that they take no serious
account of the diversity of the strategies that were open, at partieular
moments, to manufacturers and nations seeking to foster a profitable
electrical industry. Nor do they acknowledge the extent to whieh success
depended not on the adoption of some universally applicable prescription
but on the precise adaptation of strategy to the multifaceted and very
varied contexts in whieh companies had to make their decisions.
Specifically, with regard to our three slow-Iane countries, we believe
that their generally cautious attitudes towards innovation were not so
much a mark of a congenital, debilitating unadventurousness as a
calculated and often quite profitable response to the position of weakness
in whieh they found themselves. It is all too easy to forget that Britain,
France, and Italy had to take decisions, on research and innovation as on
other matters, in economic, political , and demographie circumstances
very different from those whieh Germany, for example, had to face. Even
by the late 1890s, they had a far more limited range of options open to
them than was available to one of the leading German manufacturers.
Moreover, they had to choose between those options with the experience
of a number of lessons learned the hard way. It is with some of those
lessons that the next three sections are concerned.
LEARNING THE LESSONS : FERRANTI AND THE AGE

OF A.C. IN BRITAIN
For the British, the most telling lessons were learned in the painful task of
establishing networks based on high-voltage alternating current. The
impediments, most conspieuously in London, were of a very diverse
character. As Thomas Hughes has shown, they included locallegislation
whieh was intended simply to avoid any risk of a monopoly but which led,
in practiee, to a damaging fragmentation of responsibility for the supply
of electricity to the capital. 19 In a structure that saw London provided with
no fewer than forty -nine distinct supply systems and with a baffling
diversity of voltages, frequencies , and techniques of transmission and
distribution, coherence was impossible , and the incentive to serious
investment was gravely undermined.
But there were other lessons of a more explicitly technologieal charac -
ter that demonstrated the dangers of excessive haste in the pursuit of

innovation. On this score, the cautionary experiences of Sebastian Ziani
de Ferranti quickly assumed the status of a legend. In a way that
heightened their effect, the troubles that beset Ferranti's attempts to
streak ahead of his competitors in the production and distribution of
high-voltage alternating current came hard on the heels of an exemplary
initial success. In 1886, he installed a generating station at the Grosvenor
Gallery in London in which two 700 horsepower alternators, designed and
built by Ferranti, were harnessed to the production of alternating current
at up to 2400 volts. A year later, carried along by this triumph (achieved
when he was still in his early twenties), he moved on to an even bolder
venture at Deptford on the banks of the Thames to the east of London.
At Deptford, Ferranti conceived a completely new station that was
intended to supply not just a limited locality (as his earlier station had
done) but the whole of London. The scale was grandiose, and the
technology was appropriately daring. The capacity was to be several times
greater than anything that had been handled before. The size of the 10000
horsepower alternators and the length of the transmission lines were both
unprecedented, and, with an output at 10 000 volts, the risk of break-
downs and accidents was a menacing presence. In a construction of such
novelty, in which virtually every element was designed from scratch, there
was no alternative to carrying out testing as the work proceeded on the
site. The site, in fact, became a laboratory, and experimenting became
part and parcel of the processes of construction and trial runs. It was a
precarious situation in which a combination of untried theory and
measuring and control equipment of palpable inadequacy at the voltages
in use combined to make controlled testing impossible and to force
Ferranti to try his installation by the perilous expedient of running it at
full capacity.
Inevitably, the accidents came thick and fast. A fire, burned-out
transformers, and power cuts all marked experimental failures which, in
anormal laboratory, would have been simply the starting-point for
further improvements in technique. But in the conditions in which
Ferranti was working, with the boundaries between testing and full-scale
operation blurred, they had damaging financial consequences. By 1898,
only ten years after the opening of the Deptford station, Ferranti's
company had lost half a million pounds, and it was not until 1905 that a
first, rat her paltry dividend was paid.
Ferranti himself never learned his lesson: he went on dabbling and
experimenting in a tireless quest for technological novelty for the rest of

his life. His contemporaries, however, learned quickly from what came
immediately to be regarded as a cautionary experience. The catastrophes
at Deptford stood as a warning to anyone who was attracted to untried,
innovative ventures in the electrical industry. They demonstrated the
wisdom of playing safe, and both vindicated and reinforced the caution of
investors and industrialists. Caution, in fact, became the leading
characteristic of the successful British electrical company, with
unmissable consequences for policies related to innovation. Significant
investment in research was an obvious casualty of the timorousness.
Research, after all, implied an absence of certainty and hence a shot in the
dark that any shareholder who had imbibed the lesson of Ferranti's
failure would have found unnerving.
This is not to say, of course, that the technology of electrical generating
stations and distribution networks necessarily remained static . The case of
Charles Merz, who established and administered an exemplary system for
the supply of electricity to Newcastle-upon-Tyne in the early years of this
century on the basis of a programme of steady technological improvement ,
illustrates the point very well;" It is significant, however, that Merz's
innovations were focused chiefly on design and incremental changes in
essentially well-tried practices rather than on major novelties . In the light
of Ferranti's experiences, it was a policy that might have been expected to
commend itself to other British suppliers . But in the event (for reasons we
cannot consider in this paper) the Merz model, with its principle that
technology had to be finely and painstakingly adapted to local conditions
in politics and in the economy, passed largely unheeded.
LEARNING THE LESSONS : THE ECONOMICS OF THE FRENCH

ELECTRICALINDUSTRY
In the second of our slow-Iane countries, France, it would be hard to

identify any single figure who swam as vigorously as Ferranti did against
the prevailing tide of suspicion of innovation. Here, at least by the 1890s,
there was virtually no readiness to engage in the quest for fundamental
technological novelties, and, for the historian, it is the openness of the
country to foreign machinery and practices that is most striking. 21 In some
contexts, such as the Societe Francaise pour l'Exploitation des Procedes
Thomson-Houston, the Societe Francaise d'Electricite A.E .G ., or the
Societe Westinghouse, the channel would be named subsidiaries. But,

especially where German technology was concerned, the passage into
France, for both products and senior technical and administrative staff,
would be effected more commonly through companies trading under
French names . Both Schneider at Le Creusot and the Societe Alsacienne
de Constructions Mecaniques at Belfort, for example, had agreements
with Siemens & Halske that gave them the lucrative and safe right to
exploit weIl-tried Siemens patents under varying degrees of German
supervision.
At least until the last few years before the First World War, when
nationalist sentiment intensified, there was no serious attempt to conceal
the extent of such companies' dependence on foreign technology. But the
dependence did not go unchallenged, and one of the challenges is
particularly instructive. In 1898, the promoters of what was soon to
emerge as the Compagnie Generale d'Electricite presented their venture as
a response to the invasion of French electrical manufacturing by foreign
interests. At a time when the number of truly French companies exploiting
French patents could be counted on the fingers of one hand, the CGE
declared that it would serve the national interest by pro viding an
autonomous base for the production of electrical equipment across the
whole range ." Within three or four years, however, the bid had failed.
The initial dreams of a proud technological autonomy had become a
faded memory, leaving the CGE on the well-trodden French track of
technological dependence. Now, the company's products were firmly
based on a mixture of patents leased from foreign firms, in particular the
Swiss companies of Alioth and Brown, Boveri, and on the importation of
ready-rnade or partly completed electrical machinery for erection in
France. Z3
Significantly, the departure from the declared objectives of the CGE
seems to have provoked no opposition or sense of failure. In fact , a
reading of the company's annual reports and the financial press suggests
that the new policy strongly commended itself to shareholders, especially
in the years of renewed recession between 1901 and 1903that followed the
boom of the late 1890s in which the CGE was born. As expert financial
observers repeatedly reminded their readers, it was the CGE's decision to
concentrate on the installation of weIl tested supply systems, rather than
on construction and the development of new machinery, which allowed it
to survive so successfully in aperiod when, as a young and decidedly
vulnerable company, it would otherwise have been in serious difficulty.
The state of technological dependence into which the CGE so quickly

slid remained typical of the electrical manufacturing industry in France
until the First World War. It was a dependence that left little scope or
incentive for laboratory-based research, though the scientific knowledge
and practical know-how that paid off in the on -site work of erection and
installation remained as important and, at least in certain cases, as
profitable as ever. The Compagnie Electro-rnecanique was just one of
several French companies in the sector that played to their strengths by
acquiring a reputation for the ingenuity and reliability of the mechanical
engineering on which any installation, whether for generation, traction,
or the supply of power to an industrial plant, depended. In the early
years of the century, the CEM developed its Le Bourget works as a centre
where turbines of high quality were constructed and adapted to the
alternators imported ready-made from Brown, Boveri in Baden, and it
was on this complementarity of expertise and on a judiciously
accumulated body of skill, most notably in the construction of supply
stations, that the company built its history of steady expansion and
prosperity."
The case of the CEM underlines our point that it is all too easy, in
retrospect, for the importance of the mechanical contribution to a plant
or system to be overshadowed by the scientifically more sophisticated
electrical technology with which it was allied. The innovations that
emerged from the mechanical engineering tradition (which was almost
invariably an indigenous one, often with long-established local roots)
were not ones that provoked risky new departures, nor did they or iginate
in laboratories, and, for these reasons alone, they have held little appeal
for historians. Technologically, however , they were indispensable.
Without them, the success of the hydroelectric installations in the region
of Grenoble in the two decades or so before the First World War would
have been inconceivable. Here, it was the on-site skills and creative
ingenuity of such local mechanical engineers as Neyret-Brenier which
allowed the potential of the imported German and Swiss alternators to
be realised . Such innovations as were introduced, though not negligible,
had a sure and unexciting quality that was calculated to commend them
to banks and shareholders who had been seared by the depression and
uncertainties of the 1880s and early 1890s and who were to experience
a renewed undermining of confidence when the recession returned
unexpectedly in the aftermath of the Exposition Universelle of 1900.
They were innovations calculated to satisfy the understandable and, as
we see it, dominant preference for safe and immediate returns on

investments, even if the returns were, by the standards of', say, AEG,
rather modest.
LEARNING THE LESSONS : DEPENDENCE AND ENTERPRISE

INITALY
Many of the conditions that made a substantial investment in laboratory-

based research virtually inconceivable in France also existed in Italy. Like
France, Italy about the turn of the century was overwhelmingly a receiver
of foreign technology. Here, however, the potential scope for an
independent contribution, both scientific and technological, was greater,
at least in some cases. Whereas French companies were increasingly
criticized for an excessive presence of German engineers and admin-
istrators in senior positions," in Italy technological decisions concerning
the design and construction of generating stations tended to remain in
Italian hands. In this respect, the Societä Italiana Edison, established in
1883 to supply electricity to the elegant cafes and restaurants of the
Galleria Vittorio Emanuele in the heart of Milan, was a good (and
successful) example. Yet even here the impediments to a technological
policy that might, quite reasonably, have led the firm to branch out from
installation and distribution into the manufacture of electrical equipment
and, in due course , into at least some form of independent innovation
asserted themselves. In his capacity as the company's technical director,
Colombo feit the constraints particularly keenly. When, in 1884, he
proposed that the company should establish an experimental workshop
for the production of electrical machinery and light bulbs, the idea was
rejected; the company's main financial backer, the Banca Generale,
found it too risky. 26 In 1890, the step was eventually taken, and a small
subsidiary company was set up to begin production on a modest scale. But
by now any prospect that this departure might lead on to a serious
engagement in innovation in the manufacture of generating machinery
had disappeared: foreign technology was flowing into Italy even more
freely than it had been six years earlier, so that the construction of a
commercial strategy founded on independently developed products,
which had appeared a hazardous though feasible dream in 1884, was now
quite unrealistic.
In the low priority that continued to be given to research until the First
World War, in the Societä Italiana Edison as elsewhere, Italy was not
significantly different from France. Supply companies , in particular,
tended to be heavily financed by German or Swiss banks and, by virtue of
that allegiance, were unusually open to foreign technology, while the
manufacturers and installers of electrical machinery found themselves
responding to demands that encouraged ingenuity of an adaptive rather
than a fundamentally innovative kind . Again, it is important not to
underestimate the successes that could be achieved in these circumstances,
in Italy even more obviously than in France. Italian demography and
geography, for example, were particularly favourable to the growth ofthe
hydroelectric installations which became, from the start, the country's
most notable speciality in electrical technology . It is true that France, too,
had her successes in hydroelectricity, but the distance between the main
areas of generation, around Grenoble and in the Pyrenees, and the most
populated manufacturing region around Paris remained an impediment
until an improved transmission network was in place after the First World
War. In Italy, by contrast, the sources of hydroelectricity in the Alps and
the main body of industrial and (less crucially) domestic consumers in
Milan and Turin were far closer to one another.
In this context, the potential profits and hence the incentive to reduce
the costs of production and improve the techniques of transmission were
far greater than they were in France before 1914. And it is part and parcel
of the Italian success story in hydroelectricity that this financial incentive,
far from being ignored, stimulated a major technological response. When
the Paderno generating station, near Milan, was opened in 1898, its
alternators came (as they would have done anywhere in Italy or France)
from abroad, in this case from Brown, Boveri, while the transformers
came from Ganz in Budapest. 27 But the turbines, insulating mechanisms ,
and cables were of Italian origin . Even more importantly, so too was the
overall design of the plant, which was the work of Galileo Ferraris (then
professor of electrical technology at the Polytechnic of Turin) and the
company's technical director Guido Semenza. The result was impressive:
a technologically daring installation operating at 13 500 volts and with
transmission lines extending over 33 kilometres, which at the time of its
completion in 1900 was one of the largest hydroelectric power stations in
Europe.
In this as in many other achievements of Italian electrical technology,
the element of originality was considerable, but it owed virtually nothing
to the laboratory . The contribution of the Pirelli company, which was
responsible for the cabling, was typical, For at the time, and until after the
First World War, Pirelli maintained only a small company laboratory,
devoted almost entirely to testing and product contro!. Its technological
advances in the design of electric cables for the transmission of alternating
current at increasingly high voltages were accordingly the work not of
bench-based physicists but of gifted, weIl trained engineers whose world
was that of the workshop and the site." In a pattern that had analogies in
numerous other Italian companies, the unsung heroes of Pirelli's pre-war
success were the director of the cable department, Leopoldo Emanueli,
and the firm's chief electrical engineer, Emanuele Jona, both of them
highly educated graduates of Italian engineering schools.
Another success story even more remote from the world of laboratory
research was .that of Ercole Marelli. 29 After receiving virtually no formal
education, Marelli was apprenticed to one ofthe earliest electrical firms in
Italy, the largely unsuccessful Tecnomasio, before establishing his own
modest workshop for the production of small electrical appliances in
1891. Among these appliances were the electric fans of the type that was
just beginning to be imported, for Italian domestic use, from the United
States . By 1898, the range of Marelli's products had grown remarkably,
to embrace not only domestic fans but also large and technologically more
ambitious ventilating systerns for industrial plants. The progress of the
firm was and remained a source of intense pride . By 1911, the official
company history looked back on an expansion that had taken the work
force from forty in 1897 to its present 1500 and on a vigour that had
resulted in a flourishing export market and in a manufacturing plant
equipped with the most modern machinery. Yet the only source of
technical know-how to which the publication referred was the drawing-
office, with its staff of engineers working on problems of design related to
the immediate needs of production. Equally revealing was the last
chapter, devoted to what were seen as the constraints on the future
prosperity of the firm and, more generally, of the Italian electrical
industry as a whole . Significantly, most of the constraints were
commercial ones associated with the lack of governmental support and
protection. There was no sign that either the manning or the improvement
of technical pro cesses was seen as presenting any significant difficulty.
CONCLUSION
The case histories we have reviewed are intended to illustrate, rather than
prove, our contention that the presence or absence of research
laboratories before 1914 does not, on its own , provide a satisfactory
explanation for the divergence between the countries of the fast lane and
the slow lane in electrical technology. In making this claim , we do not
underestimate the effect that the flow of scientific knowledge undoubted-
ly had on the growth of the electrical industry. We simply feel that so long
as historians focus exclusively on the contribution to the flow made by
research laboratories, and hence on major new departures, their view
must, of necessity, be a partial one . As we have argued, in this period a
main channeI for the entry of knowledge into practice remained the
essentially "low-key" route via the shop floor or engineering workshop.
When that second path to innovation is given the prominence we believe
it deserves, the boundaries between the conditions prevailing in the
countries of the fast lane and the slow lane become far less c1earthan if we
consider laboratories alone, and some existing explanations of techno-
logical success and faiIure appear exceedingly fraiI.
Plainly, in a paper of this length , it is impossible to take the argument
much further. That said, if we venture into the realms of counterfactual
history, we are convinced that the establishment of laboratories and a
serious engagement in fundamental innovative research would not have
allowed Britain, France, or Italy to make good the technological ground
they lost inexorably from the later 1880s. For laboratories to be effective ,
the context had to be one that made novelty technologically and, wherever
possible, economically rewarding. In the slow lane, the context was quite
simply not there; and no amount of investment in facilities for research
could have made up for that lack. There was, in other words, no short cut.
This disarmingly simple point carries with it a chastening message for
the historian, since its elaboration would call for a daunting degree of
contextualization and the interweaving of causes at the expense of the
monocausal explanation, We have to grapple with the fact that, in one set
of circumstances, Ferranti's technological daring led to faiIure, whereas
the research of the General Electric Company on the metallic filament and
ofSiemens & Halske on products ranging from Werner BoIton's tantalum
lamp to the high-speed telegraphic apparatus that was marketed from
1912 spawned profitable breakthroughs. SimiIarly, though in a quite
different sector , we have already referred to the divergence between Italy
and France in the development of hydroelectric power and cited the

considerations of demography and geography that must appear in any
satisfactory account.
In entering this microstructure of explanation, however, we run the
risk of evading what remains the most formidable quest ion of all. Will
contextualization necessarily take us any cioser to understanding what
were the essential differences between the group of countries that are
customarily identified as technologically and economically successful (in
the fast lane) and those that have come to be regarded as failures or also-
rans? Perhaps, for that purpose, we come back to the elusive notions of
adaptation and adjustment to the market. We suggest, in fact, that the
decisive difference lay not in attitudes towards research and in-house
innovation but in the capacity of at least the leading firms in the fast-lane
countries to offer an imaginative and creative response to the needs of the
market, whether long-felt and ciearly identified or stimulated by a new
product or application or pressure for a lowering of price. The question of
just how a company such as AEG made its response inevitably resurrects
the pursuit of the microstructure to which we have just referred. In the
case of AEG, the result would be a mosaic in which the huge size of the
company's capital (which gave the capacity for commercial risk and a
constant diversification of products), patent and tariff legislation (which
guided the international movement of goods and technologies), and the
extent of the markets for products and electric power (which set the
boundaries of opportunity) would all have their interrelated places.
One of the main methodological challenges for any of us in this field
is that this kind of multifaceted and essentially interdisciplinary analysis
only assumes real significance in a comparative setting. At least in
principle, historians have accepted this painfully obvious conciusion, and
projects and multinational ventures that reflect their acceptance are
beginning to bear fruit. But we have yet to convey the complexity of the
past to those makers of policy who are so ready to take historicallessons
out of the context that gives them meaning. The history of applied
research is an area that has been particularly misused in this way, and it
has been at least a subsidiary purpose of this paper to point to the dangers
of a selective reading of the past coloured by the preoccupations and
preconceptions of the present. We are concerned, for example, that our
contention that the efficacy of laboratory-based research before the First
World War was uncertain should not be read as necessarily implying
doubt about its efficacy in the very different circumstances of our own
day. We fully accept that, for the last century, a research-based industrial
policy, pursued in the right context, has been capable of success. Hut,
equally, it has not been a panacaea to be adopted in all places and at all
times. At least that degree of uncertainty has to be acknowledged, if
historians are to debate the significance of history with contemporaries in
search of easy and obvious solutions to the difficulties of the flagging
economies of the late twentieth century.
University oj Oxford
NOTES
I John Joseph Beer, The Emergence of the German Dye Industry, IIlinois Studie s in the
Social Sciences, No. 44 (Urbana, 111. , 1959), Chapters 7 and 8.

2 See David S. Landes, The Unbound Prometheus. Technological Change and Industrial
Development in Western Europe from 1750to the Present (Cambridge, 1969), especially pp .
323-326, and David C . Mowery, 'Industrial Research, 1900-1950', in Bernard Elbaum and
William Lazonick (eds.), The Decline ofthe British Economy (Oxford, 1986), pp . 189-222 .
J Mowery, ibid., p. 189.
4 I.C.R. Byatt , The British Electrical Industry, 1875-1914 (Oxford, 1979), pp . 150 and 166,
and Peter Hertner, '11 capitale tedesco nell'industria italiana fino alla prima guerra
mondiale', in Bruno Bezza (ed.), Energla e sviluppo. L 'industria elettrica e la Societ äEdison
(Turin, 1986), pp . 213-256.
5 The point appears on p. 172 of the English edition of the Catalogue, published as
International Exposition . Paris 1900. Official Catalogue. Exhibition of the German Empire
(Berlin, 19(0) .
6 Luigi Gabba, ' L' insegnamento della chimica nelle universitä e negli istituti superiori' ,Atti
dei JO Congresso Nazionale di Chimica Applicata (Turin, 1903), p. 64.

7 See, for exarnple, the report of the committee chaired by R. B. Haldane, published as Final
Report of the Departmental Committee on the Royal College of Science. Volume I: Final
Report with Appendix I (London, 1906). On the parallel campaign for the improvement of
scientific (as opposed to purel y technological) education and research, see Roy M.
MacLeod, 'The Support of Victorian Science: The Endowment of Research Movement in
Great Britain, 1868-1900', Minerva 4 (1971), pp. 197-230, and Peter Alter, The Reluctant
Patron. Science and the State in Britain, 1850-1920 (Oxford , Hamburg, and New York ,
1987), especially Chapters 2 and 3.
• Typical statements include : Albin Haller, 'L'industrie chimique a I'Exposition de
Chicago' , in Ministere du Commerce, de I 'Industrie, des Postes et des Telegraphes.
Exposition Internationale de Chicago 1893. Rapports publies sous la direction de M.
Camille Krantz . Comite 19. Produits chimiques et pharmaceutiques, materiet de lapeinture,
parfumerie, savonnerie (Paris, 1894), especially pp. 10- 17, and Henry Le Chatelier,
'Rapport sur les laboratoires nationaux de recherches scientifiques ', Comptes rendus
hebdomadaires des seances de l'Academie des Seiences 163 (1916), pp. 581-588 . Cf. the
laudatory comment on in-house laboratories as founts of both innovation and effective

quality control in the German chemical and metallurgical industries in Victor Cambon,
L 'A llemagne au travail (Paris, 1909), pp. 42-45 .
• International Exposition .. . Exhibition 0/ the German Empire, op . eil. (note 5), p. 172. It
is relevant to the thrust of our argument that the work of even this small proportion of the
seientifically trained employees is described not as innovation but as "examining new
inventions, testing and examining material".
10 See the report by Le Chatelier eited in note 8, above . For a statement of Le Chatelier's
views on the education of industrial seientists and engineers , see his preface to Leon GuiIIet,
L 'enseignement technique superieur a l 'apres-guerre (Paris, 1918), pp . 9-28.
11 Carlo Lacaita, 'Giuseppe Colombo e le origini deII'Italia industriale', in Carlo G. Lacaita
(ed.), Giuseppe Colombo. Industria e politica nella storia d 'ltalia. Scritt i scelti, 1861-1916
(Bari , 1985), pp. 5-86. For a typical statement of Colombo's views, see 'Le gallerie delle
macchine del lavoro e dei materiale ferroviario all'Esposizione Nazionale di Milano (1881)',
in Colombo, Scritti e discorsi scientiflci (2 vols., Milan , 1934), Vol. 2, pp, 1060-1087 .
., Anna Guagnini, 'Higher Education and the Engineering Profession in Italy , The Scuole
of Milan and Turin, 1859-1914 ', Minerva 26 (1988), pp . 512-548 .
IJ The thrust of Nathan Rosenberg's approach is expressed very clearly in the collection of
essays by hirn: Perspectives on Technology (Cambridge, 1976).

14 Leonard S. Reich, The Making 0/ American Industrial Research. Science and Business at
GE and Bell, 1876-1926 (Cambridge, 1985), espeeially Chapters 3 and 4. See also George
Wise, Willis R . Whitney, General Electric, and the Origins 0/ U.S. Industrial Research (New
York, 1985), Chapters 5-7 .
IS David Cahan, An Institute for an Empire . The Physikalisch-Technische Reichsanstalt
1871-1918 (Carnbridge, 1989); Crosbie Smith and M. Norton Wise, Energy and Empire. A
Biographical Study 0/ Lord Kelvin (Cambridge, 1989), pp . 684-698 and passim; M. Norton
Wise and Crosbie Smith, 'Measurement, Work , and Industry in Lord Kelvin' s Britain' ,
Historical Studies in the Physical and Biological Seiences 17 (1986), pp . 147-173 ; Graeme
Gooday, 'Preeision Measurement and the Genesis of Physics Teaching Laboratories in
Victorian Britain' , Brit , J. History 0/ Science 23 (1990), pp . 25-51.
16 On Regnault, see Robert Fox , The Caloric Theory 0/ Gases from Lavoisier to Regnault
(Oxford, 1971), pp . 295-302 . Kohlrausch's work as an experimental physieist is discussed in

David Cahan, 'Kohlrausch and E1ectro1ytic Conductivity: Instruments, Institutes, and
Seientific Innovation', Osiris 5 (1989), pp . 167-185; see also Cahan, An Institute for an
Empire, op . eil. (note 15), pp . 128-132. For a comment on the relations between industry
and the traditions of preeise measurement in physics, see Kathryn M. Olesko, 'Industrial
Demands and the Political Economy of Exact Experiment', read at the conference on
"Writing the History of Physics", held at St John's College, Cambridge, 3-5 April 1991.
17 On the laboratories that Siemens & Halske maintained from the earliest days of their
existence in the mid-century, and on the development laboratories of Schuckert before the
amalgamation with Siemens in 1902, see Georg Siemens, Der Weg der Elektrotechnik.
Geschichte des Hauses Siemens (2 vols., Freiburg and Munich , 1961), Vol. 2, pp. 58-59. The
very different character of AEG 's policy is brought out in Ulrich Wengenroth's recent
biographical sketch ofEmil Rathenau; see Wengenroth, 'Emil Rathenau ' , in Wilhelm Treue
and Wolfgang König (eds.), Berlinische Lebensbilder - Techniker (Berlin, 1990), pp. 193-
209, espeeially pp . 203, 204, and 207.
I. Ulrich Wengenroth, 'L 'electriflcation dans I'industrie textile', in Fabienne Cardot (ed.),
Histoire et structure economique de l 'electrification (Paris, in press), and 'Shaping a
Technological Potential: R & D in Eleetrie Drives, 1890-1930', a paper read to the historical
group of the Institution of Eleetrical Bngineers, London, in Oetober 1991 and now awaiting
publieation .
" Thomas P . Hughes, Networks 0/ Power. Electrification in Western Society , 1880-1930
(Baltimore and London, 1983), pp. 238-247 . Other aeeounts of Ferranti's early innovations
and their teehnical and eeonomic implieations include Gertrude Ziani de Ferranti and
Richard Inee, The Life and Letters 0/ Sebastian Zian i de Ferranti (London, 1934), pp. 54-
68; R. H. Parsons, The Early Days 0/ the Power Station Industry (Cambridge, 1939), pp.
21-41 ; and W. L. RandelI, S. Z. de Ferranti (London, 1943).
20 Hughes, ibid ., pp. 443-460.
2 . The point is refleeted and developed in Albert Broder, 'La multinationalisation de
I'industrie electrique fran ..aise, 1880-1931. Causes et pratiques d'une dependance', Annales
ESe, 3ge annee (1984), pp. 1020-1043.
" See the brochure announeing the formation of the company : Compagnie Generale
d'Electricite. Societe anonyme en formation . Notice (Par is, 1898), pp. 3-4.
13 The policy of the compan y emerges very clearly from the annual reports and the rich
eollection of cutting s from the finaneia l press in the file on the CGE in the Archives
Nationales (65 AQ. G.16O). The case of the CGE is discussed in Broder, 'La
multinationalisation de l' industrie electrique fran..aise', op . cit. (note 21), pp . 1031-1032.
2. The point emerges briefly from the company history , La belle histoire de la CEM (Paris,
1950), and even more strongly from the annual reports and newspaper euttings in the
Archives Nationales (65 AQ. M.118).
2S The growing resentment at the dominant position of senior German staff in Freneh
industry on the eve of the First World War is reflected in Louis Bruneau , L 'A llemagne en
France. Enquetes economiques (Paris, 1914), p. viii and passim, and in the paper s
concerning the Compagnie Generale d'Electricite de Creil (not to be confu sed with the CGE)
in the Archives Nationales (65 AQ. G.207). Feelings ran so high at Creil that they prompted
the resignation of four of the Freneh members of the Conseil d'administration in March
1912.
26 Claudio Pavese, 'Le origini della Societ ä Edison e il suo sviluppo fino alla costituzione dei
" gruppo"' , in Bruno Bezza (ed.), Energia e sviluppo, op . eil. (note 4), pp. 65-167.
27 Guido Semenza, 'L'impianto di Paderno' , Atti della Associazione Elettrotecnica Italiana
1 (1897), p. 121, and Note sui nuovi impianti della Societä Generale Italiana Edison di
Elettr icita, 1895-1898 (Milan, 1899).
2. Alberto Pirelli, La Pirelli. Vita di una azlenda industriale (Milan, 1946), and Hector
Sacchetto , History and Development of the Oi/ Filled Cable (n.p ., n.d . (1969)), Chapter 2.
2. On this case, see Vent'anni di vita della Ditta Ettore Marelli & C., Mi/ano . 1891-1911
(Milan, 1911).
H. LINTSEN . G . VAN HOOFF AND G. VERBONG
MECHANICAL ENGINEERING IN THE NETHERLANDS

IN THE NINETEENTH CENTURY: TECHNOLOGY
WITHOUT A PROFESSIONAL COMMUNITY
INTRODUCTION
Technology was an essential factor in the dramatic changes in society

which can be summed up as the Industrial Revolution. On the one hand,
it caused transformations in society, but on the other hand technology
was greatly influenced by society itself. This interaction can be shown
through the ways in which people dealt with technology and the
knowledge on which technical applications are based. For ages,
technology had been the work of craftsmen who had learned their trade in
the workshop. The major change in the nineteenth century was the rise of
professional groups, like engineers , that placed new requirements on the
education of technicians. Practical training was considered insufficient
and would have to be replaced by a more formal education at a technical
school. This process has since been referred to by sociologists and
historians of technology as the transition from a " shop culture" to a
"school culture".' At the same time, the nature of technological
knowledge changed from an entirely experience-based knowledge to an
increasingly scientific one. During the nineteenth century, technology and
science became more and more intertwined, a phenomenon which is
generally called the scientification of technology.
These two processes, the change from "shop" to "school" culture and
the scientification of technology, led the community of engineers to
become, in Layton's words, the mirror image twins of the scientific com-
munity. Today, technical education at schools holds a crucial position, for
it is here that students are trained in mathematics, the sciences and,
depending on their specialisation, one of the technical sciences; here, new
technological knowledge is passed on to the students. From this perspec-
tive a close relation between the rise of a "school culture" and achanging
body of knowledge seems evident. Nevertheless, the historical develop-
ment of the engineering profession and of the various technical disciplines
demonstrates that this was in fact an extremely complex transformation
process in which both developments were not necessarily linked.
The changes in the nature of technological kno wledge cannot, as has
155
P. Kroes and M . Bakker (eds.}, Technological Deve/opment and Science in the lndustria/
Age . 155-176.
© 1992 Klu wer Academic Pub/ishers.
156 H. LINTSEN, G. VAN HOOFF AND G . VERBONG
often been proposed, be reduced to the mere application of science. The

introduction of scientific theories and methods in technology turned out to
be a slow and difficult process that required many adaptations, new ways
of thinking and new data. Apart from that, each discipline shows its own
particular development. Neither was there always a linear relation between
the rise of technical schools and the application of scientific knowledge.
Lundgreen! has shown that the transition from "shop" to "school" before
1870 was less determined by inherent changes in technology but rather
mainly by institutional factors, like the existence of state bureaucracies.
The first professional engineers in France and Germany were employed by
these bureaucracies. A scientific education was more a criterion for the
selection of candidates for state functions than a necessary requirement for
the tasks they had to perform. In Britain and the United States such
bureaucracies were virtually absent. There, the role model for engineers
was that of the entrepreneur or independent engineer, both of whom had
received their training in practice . Before 1870 this difference in the
engineering educational system had no significant consequences for the
industrialisation in the countries just mentioned. This leads Lundgreen to
the conclusion that the (Iater) claims of these professional groups, viz.,
that they made an essential contribution to their country's industrialis-
ation, are rnainly based on ideology, not on fact. Stressing the need for a
more formal education should therefore be interpreted as being aimed at
serving professional interests . Only after 1870 did such needs gradually
become more realistic with the rise of the science-based industries.
The development of the engineering profession in The Netherlands
during the nineteenth century seems to fit in with the French and German
model. The change from "shop" to "school" is particularly apparent in
the case of the civil engineers who worked for the Rijkswaterstaat;' the
Dutch Department of Public Works. By 1800 the aspiring civil engineer
was trained by way of an apprenticeship where the emphasis was on
practical knowledge, practical experience and the personal relationship
between master craftsman and apprentice. The apprentice worked under
a senior engineer in the organisation. Meanwhile, he prepared hirnself for
his surveyor's examination by private study. The Rijkswaterstaat was set
up in 1798, and it was unmistakably a product of the process of political
and constitutional change which was taking place in The Netherlands
during the late eighteenth century . In the first ten years of its existence,
there was little division of work and only a limited hierarchy, but a
tendency towards bureaucracy existed from the beginning.
MECHANICAL ENGINEERING IN THE NETHERLANDS 157
After 1805, the Rijkswaterstaat, and thereby the profession of the civil
engineer in public service, was moulded in the French and military fashion.
It rapidly acquired a formal structure, an elaborate system of work
distribution, an extensive hierarchy and strict discipline. The engineers,
who were officers in the Rijkswaterstaat, wore uniforms and attached
considerable significance to the honour of the corps. Furthermore,
training of these engineers was transferred to military schools. In these
institutes the trainees had to conform to numerous regulations concerning
behaviour and discipline, so that an esprit de corps was fostered as weIl as
a sense of hierarchy between the cadets and the teachers, and among the
cadets themselves. The system was considered to be complete as soon as
the cadets could be separated entirely from the outside world. This stage
was reached in 1829through the boarding school facilities provided by the
newly-founded Royal Military Academy in Breda . But it was not only the
setting of the engineering education that changed, its contents had also
become subject to revision since 1805. The training became a theoretical
one with a considerable emphasis on mathematics and science. These
militarily trained civil engineers, together with the technical officers of the
army, formed the Royal Institute of Engineers in 1847. They made a
significant contribution to the changes within the Rijkswaterstaat and the
development of the Dutch infrastructure.
In this article we analyze the complex relation between the rise of a
professional community, the transformation of technological knowledge
and their effects on industrialisation. As we have taken industrialisation
as a central theme in our research , it would not be appropriate to
concentrate upon the profession of civil engineers, whose main concern
was public works. We will focus on the profession of mechanical
engineering and its relation to the development of steam technology in
The Netherlands. By shifting our attention from the civil engineers with
their evident change from "shop" culture to "school" culture to the lesser
known group of mechanical engineers, we may also be able to qualify the
implicit idea that such a professional transition holds for all engineering
sciences. Our main questions are:
1. Was there a community of mechanical engineers and does it show
traces of a change from shop to school culture?
2. Is there a relationship between the transformation of the technical
communities (from "shop" to " school" ) and the transformation ofthe
"body of knowledge", from a craft into a technical science?
3. Have professional communitie s ("shop" or "school") in the above
158 H. LINTSEN. G. VAN HOOFF AND G. VERBONG
sense really been relevant for the modernisation process in engineering

or the industrialisation process?
We should start with a short description of the Dutch technicians in the
engineering industry. But we are immediately confronted with a problem:
what kind of professional community are we dealing with here? It will
become apparent that, in the nineteenth century, there was no profession-
al community of mechanical engineers . Does this mean that the Dutch
mechanical engineers and the engineering industry did not play a
significant part in the industrialisation, and that there was no trans-
formation of mechanical engineering into a technical science? The ans wer
must consider several aspects, but in the end it is possible to argue that the
industrialisation and the development of the technical sciences in The
Netherlands were possible without the involvement of professional
communities.
MECHANICAL ENGINEERING IN THE NETHERLANDS

IN THE NINETEENTH CENTURY
One problem which appears when describing the group of Dutch tech-
nicians in the engineering industry is that of a more specific definition of
the category, In Dutch, a variety of terms were used to denote those who,
in any way, worked on machines: werktuigkundige, werktuigbouwkundi-
ge, mecanicien, ingenieur. In fact, until the start of areal school for
mechanical engineers in Delft in 1842, these terms may designate self-
made tinkerers as well as fully trained civil engineers who had gone into
machine building. Hence, even by putting together all people called
werktuigkundige we cannot claim to cover the professional group which
we are looking at in this article. In the engineering industry there were
unskilled workers, craftsmen (such as blacksmiths, fitters and boiler-
makers), the foremen, the draughtsmen, the supervisors, designers, the
director, and so on. We confine ourselves to the category of higher
technical functions, generally described as managers, engineers, heads of
departments, designers, and mechanics. We have information on the
educational backgrounds of 420 people in this category in the nineteenth
century (see Table 1).4
The majority had received their training mainly or exclusively in
practice, and had not gone through any specific training institute. They
may have attended evening classes, or qualified by means of horne study
TABLE I
Backgrounds of higher techn icians in the engineering industry during the whole of the
nineteenth century (N = 420)
Number 0/0
Practical training 308 73

Formal training 112 27
- Engineering school at Delft 58 14
- Technical education abroad 23 5
- Military school 9 2
- Technical school 6 1
- Various, or unknown 16 4
Total 420 100
and the like. Only about a quarter of the higher technicians had been
trained at some institution or other. The nature of the formal training
could differ considerably, for we may find students of civil and military
schools, schools of an intermediary and higher level, Dutch and foreign
schools, schools specifically for mechanical engineers or for other
disciplines, like building or civil engineering . Hut the practical training is
not clearly defined, either. This will become clear by looking more closely
at the types of firms where these technicians were employed.
The group of technicians worked or had worked in 160 different
engineering firms and workshops. In total, there were more than 200
companies in The Netherlands in the nineteenth century engaged in
mechanical engineering. They were spread all over the country, with
concentrations in the coastal provinces of North and South Holland. As
in most other countries, the factories and workshops differed in a number
of respects:"
- Origins: some had started as forges, others immediately started as
engineering factories, a number of others developed from trading
companies, millwrights' shops, coppersmiths and the like.
- Size: from the 1830s, The Netherlands had some large machine building
firms: Feijenoord in Rotterdam and Van Vlissingen in Amsterdam,
each employing approximately 800 people in 1847. However, through-
out the nineteenth century the smaller firms dominated. Around the
turn of the century the average company in the industry had some 100
employees. Hut the majority of companies (some 600/0) employed fewer
than 25 men.
160 H. LlNTSEN, G . VAN HOOFF AND G . VERBONG
- Independence: the majority operated independently. Sometimes a

machine shop was part of a shipyard, a railway company or an
ordnance establishment.
- Activities: machine building firms and workshops might perform a
wide range of activities. They could have foundries, forges, fitting
shops, boilermaking shops and combinations of these. Somet imes they
started with one of these activities and then extended the workshops. In
other cases, the company immediately started to carry out a large
number of activities .
- Organisation: For a long time most engineering factories and work-
shops had a horizontal structure with very little division of work .
However, during the last decades of the century significant differences
began to appear among firms where the division of labour,
organisation of work, financial management, wage structures and
administration were concerned.
With such diversity in the engineering industry it is difficult to identify
some kind of common ("shop" or "school") professional culture.
Informal networks between the mechanics and engineers may have
existed, certainly at a locallevel, but such networks do not appear to have
been very widespread, tightly knit or lasting. This is typified by the fact
that there was no national professional association until 1889. And even
when such an organisation was set up in that year under the name of
Vereeniging von Werktuig- en Scheepsbouwkundigen (Society for
Mechanical and Marine Engineers), only 62 members could be elassified
as belonging to the group of higher technicians in the engineering
industry. It is also remarkable that the first Dutch journal in the field of
mechanical engineering appeared only in 1890. Until that time there was
apparently no need for a systematic transfer of knowledge, or it had been
impossible to establish a structured exchange of information." From the
above we may conelude that there was no community of mechanics or
mechanical engineers in The Netherlands for most of the nineteenth
century. We cannot even find an informal professional community such
as that of the "mechanics" in England and the United States, who had
received their training in practice. Neither was there a more formal
professional community of mechanical engineers. The group of
technicians in the engineering industry was heterogeneous in terms of
training, experience, career and professional practice. It was fragmented
and had hardly any forms of elose cooperation. The profession of
mechanical engineer - in the sociological sense - did not exist.
However, this does not mean that there are no signs whatsoever of a
professional community. From 1860 onwards there was a nucleus among
the mechanical engineers which, by the turn of the century, would become
a community with a "school" cuIture. This core was formed by higher
technicians employed by the major engineering companies, the navy,
government services (including the State Board of Controllers of Boilers)
and the workshops of railway companies. However , the influence of this
group on mechanical engineering in The Netherlands would only be feIt in
the twentieth century.
MECHANICS, INDUSTRIALISATION AND STEAM ENGINES
The lack of any professional association then raises the question: can this
kind of heterogeneous and fragmented profession have played any role of
importance in industrialisation (in particular in connection with the supply
of capital goods) and in the transformation of engineering into a more
technical science. In answering this quest ion we will concentrate on the
construction of steam engines, and elaborate two aspects. What contri-
bution did the Dutch machine building industry make to the construction
of steam engines? How were Dutch steam engines built, what quaIi-
fications and knowledge were required, and how were these obtained?
Where the application of steam technology was concerned, The
Netherlands was not wholly dependent on foreign countries. Even in the
first half of the century, when steam was used only in a limited number of
applications, approximately half of the installations were of Dutch
manufacture. The origins of 17I steam engines are known, of which
number 85 engines were built in The Netherlands, 36 in Belgium, and 46
in other countries. 7
After 1850, when the country changed over from mainly wind and
horse power to steam power, the Dutch machine industry continued to act
as a major supplier of steam engines. Around 1880, 500/0 of all steam
engines were of Dutch manufacture. Fifteen years later the Dutch share
had increased to 55% . Around 1880, some 56% of all boilers were made
in The Netherlands, by 1910 this percentage amounted to 64% .8 When
assessing whether this is relevant or not, one ought to consider that the
Dutch engineering factories and workshops had to compete on an open
market with Belgium, England and Germany, countries with a much
longer tradition in engineering .
162 H . LINTSEN , G . VAN HOOFF AND G . VERBONG
THE BUILDING OF STEAM ENGINES
Quite independently of this assessment, one may ask how it was possible
for a country with such a fragmented profession in mechanical
engineering to be able to manufacture steam engines on such a large scale.
The answer is essentially very simple: the existence of a professional
community of mechanics or mechanical engineers was not an essential
precondition for the production of capital goods. This may become clear
by looking at the actual production of steam engines.
Building a steam engine involves a wide range of metal working
techniques:" casting, forging, machining, and joining . Cast iron and
wrought iron were the most important materials, together with other
materials such as brass and bronze, and, later in the nineteenth century,
steel. Initially timber was also used, even for some highly stressed
components, such as the crankshaft and the beam. However, in the course
of the nineteenth century the significance of wood decreased .
The process of casting involves pouring molten metal into moulds, so
that, after cooling, the required shape is obtained. Specialist skills
included the making of models, the making of moulds and the melting of
the various alloys according to precisely defined procedures to achieve the
right material properties. Apart from casting, some of the preparatory
treatment of the metal took place in the forge, in which wrought iron was
heated locally, without actually being molten. The required shapes were
produced using hammers and anvils, dies, mandreis and other tools.
Large forgings were shaped by steam and drop hammers in the second
half of the century and hydraulic forging presses by the end of the
nineteenth century .
In forging and casting, it was difficult or even impossible to give the
part the precise shape required . Especially in critical areas it was not
possible to guarantee adequate accuracy . After casting, a component
shrinks during the cooling process. Forgings were made by hand and
under visual control. No two forgings were wholly identical. Conse-
quently many components required further finishing. Many of these
operations in the nineteenth century were mainly performed using hand
tools. However, there was also an increasing use of machine tools . The
most important machines were lathes, drilling machines, planing
machines and grinding machines. Joining and finishing the various
components took place in the fitting shop which had the workbench as the
central point. Drilling, hammering, knocking, chamfering, bending,
straightening and filing were therefore the most important skills the fitter
had to possess, next to a basic sense of accuracy.
The factory or workshop building a steam engine did not necessarily
perform all these activities itself. Often it did not have a foundry, and
castings were supplied by specialist foundries . Forging, machining and
joining operations were usually performed in the factory or workshop,
even though in some cases the emphasis was on fitting, most components
being bought in. For example, the making of the cylinders with minimum
tolerances was aprecision job. But for this special tools had been
developed, which were also available to the Dutch engineering industry.
A large part of the work consisted of skilled work which required years
of experience. The operatives received this practical training on the job. A
small number had received a more formal training in industrial and
drawing schools (after 1830), technical schools (after 1860), secondary
technical schools (after 1850) and, in some cases, company training
schools. The required "theoretical" skills, such as measuring and simple
drawing, the reading of drawings and the like were taught during the
training at such schools, or were acquired through horne study. Recruiting
qualified labour was done by in-house training, by taking over
experienced workers from other firms in the same field, or by taking on
independent craftsmen, mainly from the metal working sector . In the
early period of machine production, skilled workers from abroad
dominated. They played an important role far into the nineteenth
century, but at the end of the century the large majority of skilled workers
were Dutch.
Although the very first steam engines in The Netherlands probably all
came from England and Belgium, we saw that Dutch engineering firms
soon supplied quite a substantial proportion. Before 1850, at least sixteen
companies produced steam engines .!" However, the differences were
significant. One company, P. van Vlissingen & Dudok van Heel of
Amsterdam, built 28 or more engines during this period. Some others
produced from three to ten machines, whereas the majority only
produced one or two engines. This also included some firms - not
engineering companies as such - who constructed steam engines for
themselves, such as H . Bekker of Gemert, who, in 1829, built a 4 HP
steam engine for his own cotton spinning mill.
Another example of a machine shop with a small production was J .A .
Mercx of Tilburg, who was a watchmaker by trade. Having worked for
some time in an engineering company in Aachen (Germany) , he returned
164 H . LlNTSEN, G. VAN HOOFF AND G. VERBONG
to his native town in 1835. There he started with the installation of new
steam engines and the repair of existing ones. In 1840 his affairs had
grown to such an extent that he needed a steam engine to drive lathes and
drilling machines. He built it in his own workshops, and bought a second-
hand boiler from a neighbouring cloth manufacturer. The components of
the engine were in part made by other machine shops . During the
following ten years he only built one further steam engine. Then the
production increased. During the nineteenth century eighty engines were
produced in total, including a nu mber of 150 HP compound engines.
THE DESIGN OF STEAM ENGINES
The construction of steam engines did not really pose a problem in The
Netherlands in the last century. Even a clever entrepreneur or technician
who was not involved in the engineering industry was able to rig up a
steam engine. However, this does not say anything about the quality of
steam engine engineering. This depended on the skills of the engineers , on
the workshop, and on the design of the steam engine. We now turn to the
latter, for it is the design of the installation which is at the heart of
engineering science.
During the whole of the nineteenth century - internationally speaking
- we see how the use of steam is brought to perfection, both technologi-
cally and economically. During this period there were numerous debates
and numerous discussions; '! should steam engines be vertical or
horizontal? In which way could one gain maximum advantage from the
expansion of steam? In what cases are condensers to be used, and of what
type? What is the ideal valve gear? Which design allows the minimum
steam consumption, etc.?
Such questions were the starting point for research, development and
design, and they were mainly the concern of technicians with a higher
education. What interests us here is the contribution of the Dutch
engineers and technicians to the evolution of steam technology and the
way in which they used their acquired knowledge in their designs. Is an
adequate reaction to the international technical development to be
expected from the fragmented group of higher technicians and engineers?
In view of the fact that steam engineering was developing on a broad
front, we shall confine ourselves to one aspect, namely the dimensioning
of the steam engine . First of all we will describe the international consider-
-----
~
~ "" ""\
\
I \
f
I \
I \
I I
D I J
i =---------'~
\ I
\ I
\ I
\ I
\ /
" " -, /
THE DIMENSION · QUESTION:

-- -- -- ""
Given : the required effective power Ne

Asked: D = Diameter of piston
s = Stroke
n = Number of revolut ions
Pi = Pressure of steam
10,000 x ~D2 X C
Ni X Pi
75 4
T
indicator
power
T
indicator
pressure
Talge
surtace velOCity of piston
Pj~
T,
C 2· n· s / 60
T.,
effeetive power
numberof
revolutions
stroke
Fig. 1. Dimensioning a steam engine,
ations and theories related to this question. Then we will return to The
Netherlands and investigate the Dutch position in more detail.
What is the dimensioning problem? It deals with the piston diameter,
the length of the stroke, the steam pressure and the speed of the steam
engine in order to achieve a specified power. The basic formulae for the
calculation of these dimensions are given in Fig. 1. The basis for this
calculation - the indicated capacity is proportional to the pressure in the
cylinder, the piston surface and the average piston speed - was already
166 H . LINTSEN, G. VAN HOOFF AND G . VERBONG
known in the eighteenth century. However, these formulae were not used
for design purposes until the beginning of the nineteenth century." This
was not necessary, since there were only a few types of engines,
Newcomen and Watt beam engines, for which there were tables with the
dimensions and their relation to capacity. These tables were based on
actual, working steam engines.
In the course of the last century, calculating methods became in-
Pi
Pmax. 1-----1
Pmin.
Pi = Pressure in cylinder
Pmax . = Max. pressure
Pmin. = Min. pressure
V = Volume
Vb = Volume at the beginning of expansion
V1 = Volume cylinder
Starting point
- Steam is an ideal gas
- Temperature does not change during expansion
- The law of Mariotte applies: P max Vb = p.V
_ Pmax. Vb + Pmax. v; log ~: - P min. V1

Pi =
Fig. 2. Calcu lation of Pi of a steam eng ine with expans ion (before Pambour).
creasingly important in steam engineering. The reasons were on the one

hand the increasing variety of steam engines, and on the other hand the
use of steam expansion, which made the calculation of the average steam
pressure more complicated. For the application of the basic formulae
given here, further information and additional rules were required . In
part this information was based on practical experience. For example, it
was known that the optimum ratio between piston diameter and strake
length in the case of small steam engines was 1:3, and 1:2 for larger steam
engines. When the strake length is known, then the piston diameter can
easily be calculated and vice versa. In such cases we can speak of a rule of
thumb. However, these rules changed during the nineteenth century
because of the use of new materials, new designs, new components, etc.
Other rules became the subject of profound theoretical considerations.
An important parameter, the calculation of which led to an elaborate
theories, was steam pressure .
In the development of these theories, three general periods can be
distinguished : 13
1. The " classic" theory before Pambour;
2. Pambour's theory of 1838;
3. Advanced theories after Pambour.
Starting-point
- Steam is saturated during expansion
- Temperature changes during expansion
- The law of Navier applies:
a
v=--- a , ß Practice-induced
ß+ P coefficients
P max. Vb +
Fig.3. Calculat ions with Pambour's Theor y (1838).

168 H . LINTSEN, G . VAN HOOFF AND G . VERBONG
Figure 2 illustrates the relation between steam press ure and cylinder
volume. During part of the stroke steam is supplied until volume Vb is
reached . The steam supply is then cut off, the steam expands during the
rest of the strake, and the pressure drops. The "classic" theory assurnes
that steam is an ideal gas, that the temperature does not drop during
expansion, and that Mariott's Law applies. The average pressure in the
cylinder could therefore be calculated by means of an integral (see Fig . 2).
Starting-point:
- Steam is not saturated during expansion
- Corrections on Navier and several other aspects of
steam during cyclus
- Corrections by Regnault, Rankine, Fairbairn, Zeuner, Clausius
Calculation in a German Manual (1902)
Pi = f Pmax.- f' Pmin.
i0g 0.~6 + m + 0.02 1_ 0.001 ( 1 + ö)

I _b+ m
V1
0.06+m
Fig. 4. Calculations with " advanced" theories after Pambou r.

The theory developed by the French engineer Pambour in 1838 was a

breakthrough in steam engineering (see Fig . 3). He set out from a different
set of terms of reference for the determination of the average steam
pressure. Instead of a constant temperature, he assumed that during
expansion the steam temperature would drop, that the steam
remained saturated during expansion and that Navier's Law applies. The
formulae he derived from this meant a significant improvement in the
determination of the average steam pressure.
Pambour's assumption that steam would remain saturated du ring
expansion was refuted after research by Clausius and others. They
improved his theory in this respect, and produced adjusted formulae (see
Fig. 4). These also took other elements into consideration, such as the
influence of the dead space (the volume between the valve and the piston,
when the piston has reached its dead centre), the effect of the cut-off of
the steam at the end of the stroke, and the effect of compression during
the counter-flow period. In the pressure/volume diagram in Fig. 4, the
corrections due to these factors specifically relate to the shaded area.
Various formulae were developed on this basis.
THE DESIGN OF STEAM ENGINES IN THE NETHERLANDS
New theories, methods and calculating procedures were developed in the

nineteenth century to solve the question of dimensioning. Returning to
The Netherlands, we can ask the following three questions.
1. To what extent did Dutch mechanical engineers make a contribution to
this development?
2. To what extent were they informed about the state of the art in con-
nection with this question?
3. To what extent did they apply the new approaches to the design of
steam engines?
The answer to the first question can be short : as far as we can discover,
their contribution was nil. Investigation of the specialist journals and
textbooks both at horne and abroad containing publications on the latest
developments produced nothing relating to the question dealt with here.
Neither were there any publications by Dutch authors on design in the
foreign literature. The Tijdschrift von het Koninklijk Instituut von
Ingenieurs [Journal of the Royal Institute of Engineers] gene rally showed
!ittle interest in steam engineering, and the same applies to the journal of
170 H . LINTSEN. G . VAN HOOFF AND G . VERBONG
the Koninklijke Nederlandse Akademie voor Wetenschappen [Royal

Dutch Academy of Science) and other technicalor scientific journals.
De Ingenieur, another engineering journal, took more notice of steam ,
but not of the dimensioning issue. Other journals thought theoretical
considerations too difficult for their readers . An article entitled "The
power of steam vehicles" was submitted to the editors of the Tijdschrift
van de Maatschappij voor Nijverheid [Journal of the Industrial Society).
The editors commented that they were of the opinion that "it should give
room to this subject, in spite of the rule not to publish any articles with a
subject which really belongs to mathematics or physics". 14
The answer to the second question - were the Dutch informed about
the state of the art abroad? - is more qualified . An investigation of some
20 Dutch publications relating to steam engineering in the nineteenth
century and the beginning ofthe twentieth century shows three categories.
A number of publications were published in the form of a brochure
and were intended as a general introduction for the interested layman ,
such as a popular booklet from 1854 Het stoomwerktuig populair
beschreven en verklaard [The Steam Engine Popularly Described and
Explained). These publications did not deal with the question of dimen-
sioning.
The second category consisted of textbooks and manuals intended for
use in intermediate technical education and by fitters, machine operators,
supervisors and foremen . One of these was a Praktisch handboek voor de
bepaling der afmetingen van stoomwerktuigen [Practical Manual to
Determine the Dimensions of Steam Engines) of 1860, by J .L. Nierstrasz,
or a three volume work: Het stoombedrijj. Beknopte handleiding bij de
studie van het geheeie stoomwezen voor machinisten en studeerenden
[Steam Engines. A Concise Manual for the Study of All Aspects of Steam,
for Engineers and Students) published 1921, by N.A. Imelman." Books
of this kind generally dealt with the question of dimensioning and
calculation of the power from given dimensions, but their approach was
relatively simple, with a minimum of theory . Only the basic formulae
were mentioned, and in some cases the classical theory for the
determination of steam pressure was referred to . For the rest they
provided practical rules and tables. Using the information given by these
books, the dimensioning of steam installations would have been
significantly less accurate than the existing knowledge allowed for. The
theoretical aspects of the education had low priority, as this category of
engineers was not expected to be designers.
The third category of books was intended for a higher level of technical
education and related functions, like chief engineers, directors,
supervisors, and army officers . It is typical of the situation in The
Netherlands, that only two Dutch authors in the nineteenth century wrote
books like this, namely Verdam at the beginning, and Delprat in the
middle of the century.
Verdam was a young and promising lecturer at the University of
Groningen for the application of chemistry and mechanical engineering to
the "useful arts". Although his lectures ceased in 1828 due to a lack of
interest, he remained active in mechanical engineering. Between 1829and
1837 he published a six volume work, Grondslagen der toegepaste
werktuigkunst [Foundations of Applied Mechanics), of which three
volumes deal with steam engineering." He was also an adviser to the
government in engineering matters and was responsible for the State
inspection of the safety of steam installations in The Netherlands.
Verdam had a thorough knowledge of and rich experience with steam
engineering. The completely new approach by Pambour is something that
he did not mention. He could not have known them in any case, as
Pambour did not publish his innovative ideas until 1838.
Delprat, an army officer , studied at the famous Ecole des Ponts et
Chaussees in Paris for some time . From 1821 he was a lecturer, first at the
Koninklijke Artillerie en Genieschool [Royal School of Artillery and
Engineering) in Delft . In 1829 he left this school for Breda where he
became lecturer at the Royal Military Academy, where he eventually was
appointed commander. His textbook Beginselen der werktuigkunde voor
de kadetten der artillerie, genie en van den waterstaat [Principles of
Mechanics for Cadets in the Corps of Artillery, Engineers and Public
Works), appeared in several editions, the first in 1842.11 There he
accurately followed the theoretical developments . He was weil acquainted
with the theories of Pambour, Carnot, Regnault, Clausius and others ."
It is remarkable that the one school for mechanical engineers, the
Polytechnic School which was set up in Delft in 1863, did not develop its
own textbooks on mechanical engineering . The emphasis in education in
mechanical engineering was on drawing, on reading and explaining
technical drawings and on the design of steam engines." The basis for the
calculations needed for design was dealt with in the section toegepaste
mechanica (applied mechanics). Between 1864 and 1876, Dr. L. Cohen
Stuart - civil engineer and mathematician and scientist, also director of
the Delft Polytechnic - dealt with it in a weekly two-hour lecture, for
172 H . LINTSEN, G . VAN HOOFF AND G . VERBONG
which he used a German textbook. However, according to his successor,

Professor Huet, the Delft scope was not as wide as at the Polytechnics of
Zurich and Aachen." On Huet's recommendation, significantly more
importance was attached to mechanics from 1876 onward.
Huet was the great animator in the field of mechanical engineering
education at the Polytechnic. He hirnself had become acquainted with the
subject through practice in the large engineering works of Paul van
Vlissingen and Dudok van Heel in Amsterdam. He acquired his more
mathematical and scientific training through one year of private
instruction. At the Delft Polytechnic School he taught drawing and
understanding technical drawings . This was also the angle from which he
approached steam technology . His important book on steam pumping
stations in The Netherlands consists mainly of a description of
installations and machinery . 21 This work does not deal with any
theoretical aspects, and the same applies to other publications by Huet,
which, considering his background, is not surprising.
In summing up, we may conclude that the theoretical level of steam
engineering probably varied significantly in the Dutch textbooks and at
the Dutch institutes where mechanical engineering was taught. However,
this does not appear to have prevented Dutch mechanical engineers from
schooling themselves theoretically, or from applying new theoretical
insights. In principle, any Dutch technician would have been able to
inform hirnself on the state of the art where theoretical developments in
steam engineering were concerned . Foreign literature was accessible,
although not always easily. The limited infrastructure for the exchange of
information, the foreign language of the publications and the theoretical
character of certain articles were some of the impediments . Foreign
technical textbooks were nevertheless widely distributed among machine
builders and mechanical engineers. Publications from abroad were
ordered regularly and could be found in the libraries of machine shops.
Far into the 19th century, British publications such as The Engineer and
Engineering were prominent in specialist Dutch circles. By the turn of the
century, Dutch and American journals got more attention. Large as weIl
as smaller firms subscribed to one or more technical journals. In some
instances foreign literature was also translated into Dutch, such as the
German book by J. Weisbach, Handleiding tot de kennis van het
stoomwerktuig, in 1870.2 2
Foreign technical development was accessible to Dutch engineers in
spite of their usually limited theoretical and mathematical backgrounds.
This deficiency could easily be overcome, since in the textbooks and

manuals theoretical considerations were transformed into relatively
simple calculating procedures and methods. The theoretical basis might
have been complex and only comprehensible to a specialist math-
ematician, scientist or mechanical engineer, but this certainly did not
hamper applicability of these theories. In technical books they had been
converted into tables, graphs and rules for calculation which even the less
theoretically trained mechanical engineer was able to use in practice. In
most cases the use of mathematical procedures was illustrated by means of
examples, enabling the reader to use them right away . One such example
- in a German textbook from 1902 - looks as folIows . 23
Beispiel: Für eine Eincylinder-Kondensationsmaschine mit Dampfhemd

für N; = 50 PS und Po = 5 at wird
p=4;~ = 0,15(TajII):c= 1,5(TajI);ns=30c=45; N n =33.3;

s s
I Ni
ri=O,77 und - = 1,29(Taj V);-=43.I;Ni= 1,5*43,1 =64,7;
r. c
Pi = 0,46*4 - -1,024*0,21 = 1,62(Taj.III);
75 1
F= 10000 *43,1* 1,62 =0,1996;
4"1 xd' = 1 ,03F= 0,2056;
hieraus d = 0,512 m. Wird s = 1 angenommen, so folgt
n = 30c = 45.
s
However, this is not to say that the design of the steam engine is a simple
matter. We are mere1y indicating that authors of technical books
managed to make complex information easily accessible and fit for
immediate practical use.
Whether technicians actually used the new approaches - which was our
third question at the beginning of this section - will have to remain an
174 H . LINTSEN , G . VAN HOOFF AND G . VERBONG
open question. One might expect so, but it is hard to prove, since the
formal calculation procedures are rarely found in the archives. Frequently
there are drawings, but the preparatory stages leading up to them are
difficult to reconstruct accurately . Sometimes drafts are found but the
design process employed by a mechanical engineer in the last century is
hard to reproduce.
PROFESSIONAL ORGANISATIONS AS STRATEGIC COALITIONS
In The Netherlands the mechanisation and application of steam

technology took place without the involvement of a professional
community. The dominance of the community of civil engineers and the
relatively small and fragmented machine building industry impeded the
emergence of such a community in The Netherlands untill890. Hut there
are still traces of a transformation of the " body of knowledge", as
advanced design methods based on theory came increasingly into use. The
existence of a "school"-based professional community in The Nether-
lands was not a precondition for this. Moreover, Dutch mechanical
engineers have made virtually no contributions to the development of the
principles of mechanical engineering in the nineteenth century. This may,
indeed, have been caused by the lack of a "school"-based professional
community. In general the proposition can be advanced that professional
communities are not an essential precondition for the application of the
technical sciences in industry, although they may be necessary for the
deve/opment of the technical sciences. It might be argued, however, that
this applies only to cases such as The Netherlands, i.e, to countries which
are not at the technological forefront. The Netherlands were a follower in
the nineteenth century, a small vessel drifting on a sea of knowledge, and
Dutch engineers kept a steady look-out, took soundings and scanned the
horizon .
In the twentieth century the situation changed in The Netherlands. In
mechanical engineering a "school" community has developed, which has
stimulated the establishment of powerful institutions in the field of
education, research and engineering practice. In our opinion the major
significance of a professional community is not to be found in
industrialisation or technical development. Professional organisations
have first of all a political and social significance. They are strategic
coalitions, tightly organised professional groups which in the changing
social relations are looking for recognition and power and which are
trying to create monopolies in certain areas of the labour market. This
certainly applies to the "school"-based communities and more specifi-
cally to the engineering professions, as Layton has described it for the
American engineer and Lundgreen for the engineers in Europe and
Lintsen and Disco for the Dutch engineers." So beware of the rhetoric of
the engineers. Do not believe its advocates if the professional community
considers itself to be indispensable to industrial and technical develop-
ment.
Eindhoven University 0/ Technology
NOTES
1 M.A. Calvert, The MechanicalEngineer in America, 1830-1910. Projessional Cultures in
Conjlict (Baitimore, 1967).

2 P . Lundgreen, ' Engineering Education in Europe and the U .S .A ., 1750-1930: The Rise to
Dom inance of School Culture and the Engineering Professions , Annals of Science 47(1990),
pp . 33-75.
3 H .W . Lintsen, Ingenieurs in Nederland in de negent iende eeuw. Een streven naar
erkenning en macht ('s-Gravenhage, 1980).

4 Th is table is based on the list of companies given in: W.H.P.M . van Hooff, De
Nederlandse machinejabrieken 1825-1914 (Amsterdam, 1990). This states, for each

engineering company (as far as known) , the person in control of the factory during a certain
period, his so cial origins and his training or education .
• W .H .P .M . van Hooff, In het rijk van de Nederlandse Vulcanus. De Nederlandse
machinenijverheid 1825-1914. Een historische bedrijjsstakverkenning (Amsterdam, 1990).
6 For a more detailed survey of the Dut ch mechanical engineers, we refer to W .H. P .M. van
Hooff, In het rijk van de Nederlandse Vulcanus.

7 R . Steenaard, Stoom en stoomwezen , 1824-1850, with Annex, 46 pp. (Thesis, Erasmus
Univers ity, Rotterdam, 1989).

8 Van Hooff, In het rijk van de Nederlandse Vulcanus, p . 19I.
9 Ibid. , Ch . 2.
'0 Steenaard, Stoomwezen, pp . 47-50.
11 A . den Ouden, De perjectionering van de stoommach ine voor jabrieksgebruik

(Eindhoven , n.d .), p . 2.
12 For example: Rees's Manujacturing Industry (1819-1820), in particular pp. 84 ff. ' Rules
for Determining the Proportions of Atmospheric Eng ines' .

13 P .A. Kroes, "The Role of Design in Engineering Theories' , this volume.
14 Tijdsehrift van de Maatschappij voor Nijverheid (1842), p, 470.
.. J.L. Nierstrasz, Praktisch handboek voor de bepaling der ajmetingen van

stoomwerktuigen ten dienste van werklieden, machinisten, opzigters, werkbazen, enz.
176 H . LINTSEN, G. VAN HOOFF AND G . VERBONG
(Rotterdam, 1860); N.A. Irnelman, Hel stoombedrijf. Beknopte handleiding bij de studie
van het geheeie stoomwezen voor machinisten en studeerenden (Deventer, 1921); see also:
A. Jongkees, Beginselen der stoomwerktuigkunde. Leidraad bij het onderwijs van
machinist-leeriingen (Hellevoetsluis, 1884) (2nd ed. 1891); D.J . Wagner, Beschrijving der
inrichting en werking van de stoomwerktuigen (Dordrecht, n.d.); J.C. Graue, Praktisch
rekenboek en handleiding voor het onderwijs in de stoomwerktuigkunde (Leiden, n.d.);
A.D.F.W . Lichtenbelt, Handleiding bij hel onderwijs in de beginselen der stoomwerktuig-
kunde (Rotterdam, 1905); J.A. Dittlof Tjassens, Leerboek der stoomwerktuigkunde
(Leiden, 1882) and Stoomwerktuigkunde, in four parts, with illustrations (Marine-
Machinistenschool, 1914-1915); Stoom , handleiding voor het stoombedrijf (Groningen/
Amsterdam , 1929); J .H. Harte, Volledig machinenboek ofhandboek 101 de theoretische en
praktische kennis van alle soorten van stoom en andere werktuigen en derzelver onderdeelen
(Gorinchem , 1852); E.F. Scholl, De gids voor machinisten, 4th ed, (Leiden, 1892).
• 6 G.J . Verdam, Grondslagen der toegepaste werktuigkunst, in six parts, with atlas
(Groningen, 1829-1837).
17 J.P. Delprat, Beginselen der werktuigkunde voor de kadetten der artillerie, genie en van
de waterstaal, the third edition appeared in 1861 (Breda, 1842).

• 8 It is possible to add to this category: H.W . Schroeder van der Kolk, Hel behoud van
arbeidsvermogen bij de stoomwerktuigen (Leiden, 1861). The author comments on

Pambour's and other's theories . However, the publication relates to a presentation to the
Natuurkundig Gezelschap te Utrecht [the Physical Society of Utrecht] and was not directly
addressed to mechanical engineers or designers of steam engines.
19 Handelingen der Staten-Generaal. Bijlageen Register, 1868-1869. Report on the state of
the Higher, Intermediate and Lower Education 1867/1868, p. 1412.

20 A. Huet, De Delftsche Akademie en de Polytechnische School; Verspreide schetsen en
studien (Purmerend, 1880).

2 . A. Huet, Stoombemaling van polders en boezems ('s-Gravenhage, 1885).
22 J. Weisbach, Handleiding tOIde kenn is van het stoomwerktuig, translated by G. Kuyper
(Amsterdam, 1870). This textbook goes into detail about the theories of Pambour and
others . Other foreign publications translated into Dutch: Chr. Cremer, Hel stoomwerktuig
(Rotterdam, n.d.): simple handbook for direct technical training, G. Pipijn and A. van den
Steen, Leerboek der werktuigkunde, stoomketels & stoommachines met praktisch
onderricht o ver het vervaardigen en behandelen der gewone stoomwerktuigen (Gent, 1876);
textbook for direct teaching at technical colleges (The fourth ed, appeared in 1897); J.
Bourne, Leer- en handboek der stoomwerktuigkunde (Nieuwediep, 1858); practical
handbook for technical operatives and companies, Bernoulli's Vademecum , practisch
handboek voor berekeningen, dagelijks voorkomende in de bouw- en werktuigkunde,
originally translated by J.G. van Gendt jr ., revised by G.J. W. de Jongh (Amsterdam, 1884).
23 Des Ingenieurs Taschenbuch (Berlin, 1902), p. 822.
24 E. Layton, The Revolt of the Engineers; Social Responsibility and the American
Engineering Profession (Cleveland, 1971); H.W. Lintsen, Ingenieurs in Nederland; C.

Disco, Made in Delft, Professional Engineering in The Netherlands 1880-1940, Ph.D . thesis
(Amsterdam , 1990).
EDA KRANAKIS
HYBRID CAREERS AND THE INTERACTION

OF SCIENCE AND TECHNOLOGY
INTRODUCTION
Ideas about the science-technology relationship have gone through an

important evolution over the last decades. In the 1950s and 1960s the
dominant model portrayed technology as appIied science. This model
assumed a hierarchical, almost parasitical relationship between science
and technology . It assumed that technological development followed and
was dependant upon paths of scientific change, whereas science followed
its own, internal line of development, largely independent of technology .
By the 1970s,the Iimitations and deficiencies of the appIied science model
had become increasingly apparent. Through the efforts of Edwin Layton
and others, an alternative model was developed which has since gained
wide acceptance . This model portrays science and technology as two
distinct, but interacting communities, each with its own traditions, goals,
and values, and its own body of knowledge and technique . The two
communities borrow from one another, but on their own terms , generally
transforming the borrowed knowledge to adapt it to different ends. In
Layton's words:
This model [of technology as applied science] ... assumes that science and technology
represent different functions per formed by the same community. But a fundamental fact is
that they constitute different communities, each with its own goals and systems of values .
They are , of course, similar in that both deal with matter and energy, But these similarities
should not be overstated. Each community has its own social controls - such as its reward
systern - which tend to focus the work of each on its own needs . These needs determine not
only the objects of concern, but the "language" in which they are discussed. These needs
may overlap; but it would be surprising if this were a very frequent occurrence. One would
expect that in the normal case science would beget more science, and technology would lead
to further technology. I
DeIineating the boundaries and distinctions between science and

technology in this way has had important historiographical benefits. It
has had the benefit that technology is analyzed on its own terms rat her
than as a subordinate adjunct to science. It has also been instrumental in
177
P. Kroes and M. Bakker (eds.), Technological Development and Science in the lndustrial
Age, 177-204.
178 EDA KRANAKIS
opening up a new domain of research that focuses on the nature and

varieties of technological knowledge in relation both to scientific
knowledge and to innovation. This research has shown that technology -
even advanced technology - is not necessarily Iinked hierarchicaIly to
science, and further , that technology has a cognitive dimension and an
inteIlectual history that are not merely subsets of those of science,?
Nevertheless, it remains true that the social and cognitive realms of
science and technology have never been entirely distinct, and in fact seem
to have become increasingly intermeshed and interdependent over the
course of the last two centuries. And despite its other advantages, the
"two communities" model, as it has generaIly been interpreted, does not
fuIly aIlow for this phenomenon. I believe that to get new insight into the
nature and evolution of the intermeshing of science and technology , it is
helpful to adopt a model that views science and technology not as distinct
communities that intermittently interact, but rat her as overlapping realms
of social activity ,? The model I propose assumes overlap in the
communities of science and technology as weIl as in organizational
structures, bodies of knowledge, traditions of practice, and value and
reward systems. The two realms are distinguished in this model by the
differences in their principal social activities: while the primary activities
of science are the creation, screening, codification, and dissemination of
official public knowledge about the natural world, technology is first of
aIl oriented toward the production and maintenance of society's material
infrastructure."
Modelling science and technology as overlapping worlds has several
advantages ." First , by caIIing attention to the areas of intermeshing of
science and technology, this model encourages us to examine more
carefully how these domains are Iinked. Second, the model assumes that
the area of intersection between the worlds of science and technology is
multi-dimensional. As suggested above, it may be defined in terms of
intersecting organizational structures, bodies of knowledge, traditions of
practice, communities of practitioners, etc.. Each dimension suggests a
distinct line of inquiry or focus for research. A third advantage of the
model is that it is consistent with a dynamic view of the relationship
between science and technology. It assumes that the nature and extent of
their intersection is fluid and changing, varying according to time, place,
and discipIine, and dependant upon historical and contextual
circumstances. In other words, the nature and extent of this intersection
is assumed to be sociaIly constructed . Finally, this model allows for the
THE INTERACTION OF SCIENCE AND TECHNOLOGY 179
possibility that science and technology may not retain distinct boundaries
and separate identities. It suggests, indeed, that we should expect to find
a phenomenon of hybridization occurring, in which a merging or blending
of traditions, practices, knowledge, and values from the worlds of science
and technology takes place.
The present paper focuses on this phenomenon of hybridization at the
level of individual , hybrid careers. The idea of a hybrid career was hinted
at by Edwin Layton in his article, 'Mirror Image Twins: The Communities
of Science and Technology in Nineteenth Century America' . There, in a
brief paragraph, he made the comment that information exchanges
between the worlds of science and technology depend upon individuals
who straddle both worlds:
For information to pass from one community to the other often involves extensive
reformulation and an act of creat ive insight. This require s men who are in some sense
members of both communities. These intermediaries might be called " engineer-scientists"
or " scientist-engineers" , depending on whether their primary identification is with
engineering or science. Such men playa very important role as channels of communication
between the communities of science and technology."
The aim of this paper is to take this observation further, and to explore
hybrid careers as a means to learn more about how the domains of science
and technology are interconnected. The term itself, "hybrid career" is
intended to suggest two methodological criteria for such an analysis.
First, "career" is intended to emphasize the idea of a path over time; it
emphasizes that we need to look at how lifetime career paths wend their
way back and forth between the domains of science and technology. Such
an approach is important because the structure and trajectory of a career
path helps to shape the activities that are carried out in each domain, and
thus influences the way these domains are interlinked.
Second, the term "hybrid" is intended to suggest that an individual
need not retain a primary affiliation with either the scientific or
technological community over the course of a career (as Layton implies).
The term further cautions us against the common practice of designating
someone's entire career as "scientific" or "technological" based on a
segment of their career path . Too often such designations lead to the
assumption that any crossing of boundaries by an individual from one
domain to the other is a temporary aberration or sideline with little
bearing on the person's "principal" line of creative work . In contrast, the
idea of a hybrid career suggests that such crossing of boundaries may be
180 EDA KRANAKIS
a structural characteristic of the career which shapes the activities carried

out in each realm.
This paper makes two points about hybrid careers which concern,
respectively, their social and organizational roots, and their influence on
practice and knowledge production. The first point is that hybrid careers
are fostered by interorganizational networks that create structural links
between the worlds of science and technology. The second point is that
hybrid careers lead not only to the diffusion of knowledge and
information between the two worlds, but also to the creation of what may
be termed hybrid repertoires of practice. Traditions of practice in both
science and technology depend upon particular stocks of knowledge and
particular skills, methods, concepts, values, ideologies, priorities, etc .
Individuals (and also groups of scientific or technological practitioners)
have specific repertoires of these elements. What happens in the case of
hybrid careers is that new repertoires are formed which combine elements
from the two realms.
These two general points will be illustrated in greater detail by looking
at three examples of hybrid careers, seeing how each was linked to a
particular interorganizational network and how each resuIted in a
merging or hybridization of knowledge, practices, and values from the
worlds of science and technology. The examples include: (1) Claude Louis
Navier (1785-1836), an engineer with the French Corps des Ponts et
Chaussees who also did research in mathematics and theoretical
mechanics; (2) William Thomson (Lord Kelvin) (1824-1907), who
combined a distinguished scientific career with a lifelong involvement in
engineering and technological innovation; and (3) Vilhelm Bjerknes
(1862-1951), a theoretical physicist who became a practical meteor-
ologist.?
I. CLAUDE LOUIS NAVIER
Navier was trained as an engineer and served in the state engineering corps
throughout his career. As a student, he attended the Eco/e Po/ytechnique
and the Eco/e des Ponts et Chaussees . After graduating in 1807, he
became a memb er of the Corps des Ponts et Chaussees, stationed in Paris,
with the rank of ingenieur ordinaire. By 1822 he had attained the rank of
Ingenieur-en-chef:" Over the years Nav ier was involved in many practical
engineering projects, including the design and construction of several
bridges. There can be no doubt that he had a deep commitment to the

engineering profession. Indeed, he saw hirnself as following in the
footsteps of his uncle, Emiland Gauthey, a distinguished member of the
Corps des Ponts et Chaussees who had raised Navier from the age of
fourteen. In Navier's words, Gauthey bequeathed to hirn his "heritage as
engineer". The character of Navier's career came to differ significantly
from his uncle's however. 9 For whereas his uncle's career remained solidly
rooted in the realm of practical engineering, Navier's came increasingly to
move between the worlds of science and engineering. 10
Navier's links to the world of science can be seen in the social relations
he established within the scientific community, and in the honours which
the scientific community bestowed upon hirn. These honours included,
most notably, his nomination to the Societe Philomatique in 1819 (a
major stepping stone to the Academie des Sciences), 11 his election to the
A cademie des Seiences in 1824, and his appointment in 1831 to what had
been Cauchy's chair in analysis and mechanics at the Ecole Poly-
technique. Socially and intellectually, Navier became closely linked with
a group of mathematicians including Poisson, Cauchy, Sophie Germain,
and most importantly, Fourier. The latter bequeathed his papers to
Navier, who subsequently prepared a posthumous edition of Fourier's
Analyse des Equations Determinees (Paris, 1831), with an historical
introduction on Fourier's research in this area."
The organizational context in which Navier worked helps to explain his
dual orientation. In Navier's environment, the worlds of science and
technology were structurally coupled through an interorganizational
network that linked the major engineering schools (the grandes ecoles,
including the Ecole des Ponts et Chausseesi and the state engineering
corps, with the Academie des Sciences. This structure was something new
in Navier's time; it was most immediately the product of the creation of
the Ecole Polytechnique and the reorganization of the state engineering
schools in 1799, in the wake of the French revolution.
Prior to the French revolution, engineers in the Corps des Ponts et
Chaussees (like Gauthey) were trained only at the Ecole des Ponts et
Chaussees. The education offered was something like an apprenticeship;
there were no formally organized courses. Experienced students helped to
tutor the newcomers, and summers were spent visiting various
engineering projects. The level of the mathematics education provided
was fairly rudimentary. Also, there were no direct links between the Ecole
des Ponts et Chaussees and the Academie des Seiences at that time."
182 EDA KRANAKIS
By 1800, the situation was quite different, however. Now anyone

wanting to enter the Corps des Ponts et Chaussees had to complete five
years of formal education, two years at the Ecole Polytechnique, followed
by an additional three years at the Ecole des Ponts et Chaussees. The
emphasis of this education was heavily upon mathematics and theoretical
mechanics. In fact the education gave bright students the ability to move
very rapidly to the forefront of research in these areas. Equally important,
the Ecole Polytechnique was now directly linked not only to the Ecole des
Ponts et Chaussees but also to the Academie des Sciences, because the
committee that oversaw the curriculum of the Ecole Polytechnique
officially included members of the Academie des Sciences." Closer links
also evolved between the Academie des Seiences and the engineering
corps. Especially within Paris, members of the two groups moved in the
same social circles, often attending the same salons, for exarnple."
Members of the engineering corps also held a significant percentage of the
positions in those sections of the Academie des Seiences most relevant to
engineering, such as the mechanics section."
Shaped within this context, Navier's research interests moved from
problems and topics rooted in the practice of engineering, toward an
additional interest in problems of concern primarily within the scientific
community. Navier's early publications were primarily within the realm
of engineering. These included, most notably, an edition of Gauthey's
works on bridges and canals (for which he prepared extensive notes and
additions), and the preparation of thoroughly revised editions of two
classic, eighteenth-century French engineering treatises by Belidor:
Science des Ingenieurs and Architecture Hydraufique . Navier's additions
to these texts focused on theoretical engineering questions: stability of
arches, theory of friction, etc. They represented a transformation in the
analysis of these problems (as given in the original texts) by bringing to
bear upon them more sophisticated theories and techniques from
mathematics and theoretical mechanics.
Many of Navier's later publications (after 1820) also dealt with
engineering questions, but he also began to focus on research questions
emanating directly from the scientific community. For example, at a
meeting oftheAcademie des Seiences in 1808, E. F. Chladni had given an
experimental demonstration of the modes of vibration of flat plates.
(These vibrations were made "visible" by covering the plates with sand.)
Following this demonstration, the academy proposed a prize for an essay
that would develop the mathematical theory of the phenomenon . The
mathematician Sophie Germain was eventuaIly awarded the prize in 1814,

but with the explicit reservation that she had not fuIly resolved the
problem . At that point Poisson took up the problem, and published his
own analysis of it shortly thereafter. Navier became interested in the
question as weil, and in 1820 he presented a memoir to the academy in
which he used a different approach and different techniques (including
the recently developed mathematical technique of Fourier analysis) in
order to extend Poisson's results. Navier further generalized his own
results in a memoir the following year, a memoir that is generaIly
acknowledged as having been a major step toward the development of a
general mathematical theory of elasticity. Navier 's work was then
immediately taken up and extended by Cauchy, and it provided an
impetus for further work by Poisson and others as weIl.17
Navier thus had a kind of dual career : one in which he functioned
alternately as a scientist and an engineer. Yet he did not keep these
activities completely separate. Within his engineering career he developed
a repertoire of knowledge, techniques , and values that combined elements
from both his scientific and technological experience. For example,
Navier's knowledge of mathematics and theoretical mechanics became a
fundamental element of his repertoire in engineering. A good example can
be seen in work he did on suspension bridges in the early 1820s. His
concern with this technology stemmed from an official request by the
administration of the Corps des Ponts et Chaussees in 1821 to study and
report on the progress of suspension bridge construction in Britain.
Navier subsequently made two trips to Britain, during which he gathered
source materials, spoke with leading civil engineers, and carefuIly
examined the most important suspension bridges that had been built
there. The report which he subsequently presented to the corps' director,
however, was no ordinary, descriptive engineering report, but rather a
book-Iength treatise which contained a 250-page mathematical theory of
suspension bridges . The capstone of the treatise was an analysis of cable
vibrations in suspension bridges, based on what is now termed Fourier
theory (and which was then at the cutting edge of mathematics research).
The way Navier conceptualized the cable vibration problem made it
analogous to the vibrating string problem (a c1assic mathematical
problem) . Navier's subsequent election to the Academie des Seiences was
partly in recognition of his theor y of suspension bridges, but he also used
the theory as a practical guide in designing a suspension bridge over the
Seine, in Paris."
184 EDA KRANAKIS
Navier's engineering repertoire included not only scientific knowledge

but also values rooted in his scientific career. Perhaps most important was
a belief in the supremacy of mathematics. When Navier competed for
election to the Academie des Sciences, he wrote to a colleague:
As to the election, I forbid myself, as you can imagine, any remark about the person or the
work of my competitors.... In confining myself thus to general considerations, I cannot help
but hope that longer consideration will perhaps bring a person who has given great service
to rnathematics, on the belief that this species of knowledge is that which it is most irnportant
to encourage and to honour, and which will always const itute the most solid foundation for
the glory of the Academy. ' 9
Within the context of Navier's engineering career, this belief in the

supremacy of mathematics was translated into a belief in the possibility
and desirability of using mathematical theory as a primary foundation for
engineering practice . Starting from this premise, Navier developed a
mann er of problematizing technology in which practical engineering
problems were transformed into theoretical problems accessible to
mathematical analysis. His treatise on suspension bridges, for example,
transformed the problem of how to design a safe and economical bridge
into the mathematical problem of analyzing the statics and dynamics of
an idealized cable.
Navier also developed specific techniques for using theory to guide
practice , and these techniques likewise became part of his engineering
repertoire. Three techniques in particular may be mentioned . The first
was to use mathematical analysis to specify the theoretical limits within
which certain phenomena would operate in a technological structure. The
second was to determine the relationships between important parameters
or functional cbaracteristics of a structure - that is, to determine how they
varied relative to one another." In the case of the suspension bridge, for
example, Navier determined how the deflections in a cable produced by a
given load would vary relative to the cable's span (i.e. the horizontal
distance between its two ends), assuming a constant sag-span ratio. Tbe
third technique he employed was that of "selective complexification" . In
order to be able to model technological structures mathematically, Navier
bad to devise simplified, idealized characterizations of the structures.
Sometimes the degree of idealization necessary to make the mathematics
tractable was so great that the resulting analysis could give very little
information about many aspects of a structure's real behaviour. Navier's
technique of selective complexification was intended to get around this
problem. What he did was, first, to develop a very abstract and idealized
mathematical model of the technology he was investigating, and then to
add complexifying hypotheses, one at a time, so as to gain more accurate
information about selected features of the system's behaviour. To return
to the example of the suspension bridge, Navier first modeled it simply as
a massless, perfectly flexible, inextensible cable loaded uniformly along
the horizontal. He then progressively modified his analysis to account for
the influence of the following factors on the cable's equilibrium (each
factor being considered separately):
(1) the actual weight ofthe cable and ofthe hangers (which link the deck,
or roadway, to the cable);
(2) additional concentrated or distributed loads on the cable;
(3) the cable's actual elasticity (i.e. the elasticity of iron) ;
(4) the expansion or contraction of the cable due to temperature changes;
(5) the kind of tower-cable system adopted.
Navier's hybrid repertoire clearly influenced the kind of knowledge he
produced. For example, he contributed significantly to the growth and
development of the theory of elasticity and the theory of structures, but
hardly at all to the elaboration of experimental, design, or production
techniques in engineering (nor did he ever take out a patent.) Navier was
part of an engineering elite in France, and the interorganizational network
of which he was apart, while it brought hirn into close contact with French
scientists, tended to isolate hirn intellectually and socially from engineers
and entrepreneurs in the private sector . The kind of knowledge Navier
produced reflected this situation. He was not one to devise simple,
accessible rules or formulas that the technologists of his day could readily
apply. To read his work at all presupposed a knowledge of mathematics
and theoretical mechanics that was far beyond the ken of most of them .
Indeed, when considered relative to the standards of most technologists of
his time, the kind of knowledge Navier produced appeared esoteric and
not immediately relevant to practice.
2. WILLIAM THOMSON (LORD KEL VIN)21
Like Navier, William Thomson had a "dual" career in the worlds of

science and technology, yet as we will see, it differed in character from
Navier's. Thomson attended Glasgow University and then studied at
Cambridge (St. Peter's College), where he graduated in 1845 as second
186 EDA KRANAKIS
wrangler and Smith's Prizeman. Following his graduation, he spent some

months in Paris, attending lectures and working in the laboratory of
Victor Regnault. He also took up (until 1852) the editorship of the
progressive, research oriented, Cambridge Mathematical Journal. In
1846 Thomson was appointed professor of Natural Philosophy at
Glasgow University, where he remained until his retirement. In 1851 he
was elected a Fellow of the Royal Society, and in 1890 he became its
president. Throughout his career, Thomson published extensively, and
his scientific investigations were wide ranging . For example, he made
important contributions to the development of mechanics, thermo-
dynamics, and electromagnetic field theory, and he contributed to
theoretical debates in geology and cosmology. He was recognized
internationally as one of Britain's leading scientists, and upon his death
was buried alongside Isaac Newton in Westminster Abbey.
In addition to a distinguished scientific career, Thomson also became
deeply involved - socially, intellectually, and professionally - in the world
of technology and engineering. He became a member of the Society of
Telegraph Engineers (later renamed the Institute for Electrical
Engineers), and served as president ofthe society twice, in 1874and 1889.
He became a scientific or engineering consultant to many companies. He
served as a member of the Board of Directors of the Atlantic Telegraph
Company, and was knighted for his contributions to the laying of the
Atlantic Cable. He served on government committees relating to
technology, such as the House of Commons Select Committee on Electric
Lighting. He became a successful inventor, obtaining patents in areas such
as telegraphy and navigation. In the case of his telegraphy inventions,
Thomson formed a Iong-term business and consulting partnership in the
1860swith two engineers, Fleeming Jenkin and C.F . Varley, in which the
three pooled their respective patents and shared the profits that accrued
from them . In another case he formed a company to manufacture and sell
navigational equipment based on his patents in that area.
Thomson's career path within the worlds of science and technology
was shaped by an interorganizational network that promoted contacts
between academic scientists on one side, and inventors, engineers, and
entrepreneurs on the other. A complete analysis of this network would
require looking at diverse institutions and groups around Britain, such as
the British Association for the Advancement of Science, government and
military advisory committees, and the Institution of Electrical Engineers.
For the purposes ofthis paper, however, let me focus only on two import-
ant institutions within Thomson's immediate environment: the Glasgow

Philosophical Society and Glasgow University .
As Smith and Wise have emphasized, the Glasgow Philosophical
Society was notable for the way it brought academics into regular contact
with members of the industrial community. Thomson was elected to the
society in 1846and participated in its affairs throughout much if not most
of his career . There he became interested in problems relating to
engineering and technological development - for example, problems
concerning ship performance, the design and efficiency of prime movers,
and the efficient production and use of materials. And there he
established lifetime friendships with engineers and industrialists like
Walter Crum, a cotton manufacturer, and J .R . Napier, who had the
position of naval architect in his father's shipbuilding firm ."
Glasgow university fostered contacts between scientists and engineers
by bringing engineering into the university : achair of civil engineering and
mechanics was established there in 1840, one of the first such chairs in
Britain;" Thomson interacted extensively with the first two holders of this
chair, Lewis Gordon and William J . M. Rankine (both of whom were also
active in the Glasgow Philosophical Society) . For example, it was through
Gordon that Thomson first obtained a copy of Carnot's theory of heat
engines, in 1848, and when he subsequently prepared an article on
Carnot's theory, he sent it to Gordon for comments prior to publication.
Thomson's friendship with Gordon continued after the latter resigned his
chair in 1855to set up his own engineering firm. Both were involved in the
Atlantic Cable project and corresponded on that subject, (Gordon was
involved with a firm that manufactured parts of the cable This firm also,
through Gordon, agreed to patent some of Thomson's telegraphy
inventions.) And it was Gordon who introduced Thomson to Fleeming
Jenkin, with whom Thomson later established a business partnership.
Both Gordon and Rankine, but especially Rankine, helped to shape
Thomson's ideas in mechanics and thermodynamics. Thomson had
discussions and correspondence with Rankine on these topics and paid
close attention to Rankine's papers. Thomson also applied for a patent
with Rankine."
The third holder of the chair of engineering at Glasgow (from 1873)
was none other than Thomson's brother, James Thomson. James
Thomson had studied engineering under Gordon at Glasgow, after which
he served an apprenticeship under the renowned machine builder,
William Fairbairn. In 1854 he took up the chair of engineering at Queen's
188 EDA KRANAKIS
College, Belfast, where he remained until accepting the Glasgow chair.

Heimholtz wrote about Thomson's brother that "he cares for nothing
except engineering, and talks about it ceaselessly all day and all night, so
that nothing else can be got in when he is present. "25 It is clear that J ames
Thomson played an important role in turning Thomson's attention
toward engineering questions and perspectives, and in reinforcing
Thomson's social and intellectual ties with his engineering colleagues.
Jam es Thomson also shaped the scientific thinking of his brother
through extensive discussions and correspondence. In the words of Smith
and Wise, "Jam es played a central role in the creation of his brother's
thermodynamics, energy physics, cosmology, and matter theory."26 For
example, one of Thomson's major scientific accomplishments was his
reformulation of mechanics. Whereas classical Newtonian mechanics was
built around the concept of force, William Thomson restructured
mechanics around the concepts of work and energy. The concept of force
was reduced to a secondary status and reinterpreted in macroscopic,
engineering terms. His brother James's role in this process was at least
two-fold. First, it was through James (and Lewis Gordon) that William
Thomson first began to give serious attention to the concept of work.
Second, William Thomson's reformulation of Newtonian dynamics
specifically involved reformulating Newton's third law around the
concept of work, and there is good evidence that Jam es was the initiating
force behind this reinterpretation.
As Smith and Wise repeatedly stress, the knowledge, perspectives, and
values that Thomson acquired through his associations and experiences in
the world oftechnology did not remain peripheral to Thomson's scientific
work, but in fact profoundly shaped its style and content. He developed
an ideology which saw technology as "the life and soul of science" .27
More specifically, he came to believe that the best way to understand and
conceptualize nature was not through abstract mathematics but through
the process of controlling nature, of harnessing it to achieve practical
tasks. This belief shaped Thomson's scientific methodology in at least two
ways, In the first place, he frequently used technological models and
theories as guides to physical intuition and theory. To quote Smith and
Wise, Thomson accepted "a methodology that grounded mathematical
physics in steam-engines, vortex turbines, and telegraph lines. "28 His
theoretical views about electricity, for example, were modeled directly
around his telegraphy theory. The latter analyzed electricity as if it were
a fluid, with conduction represented as a "flow" of electrical density. The
fluid analogy subsequently became the foundation upon which he tried to

explain a broad range of electrical phenomena and indeed to build a
general theory of electricity.
Thomson's belief that control of nature leads to scientific under-
standing also shaped the standards he used to evaluate scientific theories,
and to distinguish "good" from "bad" science. To give one example,
Thomson translated his reliance upon physical or technological models
and analogies into a standard for the evaluation of theories : theories that
could not be comprehended in terms of such models were not acceptable.
As one of Thomson's colleagues explained:
According to the standard that Thomson had set up it is not sufficient to obtain an analytical
result, and to reduce it to numerical computation, every step in the process must be
associated with some [physical) intuition, the whole argument must be capable of being
conducted in concrete physical terrns."
Applied in the case of electricity, this standard led Thomson to the view
that the concreteness and proven efficacy of his telegraphy theory
(physically embodied in the Atlantic cable) was a valid justification for
accepting a theory of electricity grounded in the same model. In contrast,
he rejected Maxwell's theory of electromagnetism, in part because he
could not associate it with any concrete, physical intuition or model.
Another crucial element in Thomson's scientific repertoire, which also
reflected his technological experience, was a commitment to precision
measurement. He saw precision measurement taking on an increasingly
important role in technological development. Time and again, either
through his own experience or through his knowledge of the experiences
of his engineering colleagues, he saw thousands of dollars being saved or
lost through measurement or inattention to measurement; he saw
companies and technological projects succeed or fail, seemingly directly
in relation to their directors" commitment to precise measurement. To
give one example, Thomson's telegraph theory showed quantitatively the
retardation of the signal that would occur in underground and underwater
cables, relative to their length. The theory further showed how such
effects could be minimized or compensated for by increasing the diameter
of the wire and the thickness of the insulation in relation to the cable's
length . Yet Thomson realized that to be able to design cables in accord-
ance with this theory, precise measurements were required, for example
on the resistivity of copper and the specific inductive capacity of gutta-
190 EDA KRANAKIS
percha. He carried out some measurements on the resistivity of copper

himself, and they made clear that the performance, or alternatively, the
cost of underwater cables would vary significantly, depending upon
which supplier provided the copper. In this way, Thomson showed that
precise measurements were a matter of immediate practical and financial
concern for "share-holders in submarine telegraph companies", 30
Given Thomson's view of measurement as a crucial factor in achieving
control of nature, and given his equating of control with understanding,
it is hardly surprising that he also came to regard measurement as crucial
to scientific understanding. In Thomson's view, control, measurement,
and scientific understanding formed an inseparable triad, in which the
quest to harness nature to achieve practical tasks led to more sophisticated
traditions of measurement, which in turn led to a deeper understanding of
nature." The laboratory he built up at Glasgow university, which became
"the first professional research and teaching laboratory of natural
philosophy in Britain, "32 was a testimony to this belief. It focused
particularly upon instrumentation and "precision measurements of the
properties of matter':", and was oriented not only toward teaching and
professional scientific research, but also toward technological research .
Indeed much research there was carried out directly in relation to
Thomson's technological and entrepreneurial activities. (His tests on the
resistivity of copper provide one example.)
Thomson's commitment to knowledge grounded in precise measure-
ment was such that he rejected theories that contained unmeasurable
entities, the most notable example being Maxwell's theory with its un-
measurable "displacement current". For Thomson, empirical measure-
ment set the limits for theorizing: mathematical theories could never be
allowed to go beyond the analysis of empirically known and directly
measurable entities or phenomena. Maxwell's theory flouted this rule,
and Thomson accordingly regarded it as a dangerous form of mathema-
tical idealism.
One other element of Thomson's scientific methodology that needs to
be mentioned was his focus on the analysis of work, a focus which again
derived from his engineering experience. The first exemplification of this
approach occurred in 1845, in a theoretical analysis of an electrostatic
system consisting of two interacting, conducting spheres. Thomson
analyzed the problem by calculating the work done on (or by) the system
during an infinitesimal change in the distance between the two spheres.
This was a radical departure from established approaches in electrical
theory, which focused on forces acting between point masses." In effect,

Thomson treated the system as an engine, and analyzed it the way
engineers analyzed heat engines and other prime movers, i.e, in terms of
the total work or energy going into and leaving the system. This approach
allowed hirn to derive a simple expression for the attraction or repulsion
of the spheres, a task that had earlier defeated Poisson, who analyzed the
problem in terms of point forces. Thomson subsequently extended his
new approach to many other areas of scientific inquiry. It was reflected,
for example, in his reformulation of Newtonian mechanics, and he even
used it as the basis for carrying out precision measurements and for
designing measuring instruments .35
3 . VILHELM BJERKNES 3 •
Unlike Navier and Thomson, Bjerknes did not have a dual career in the
sense of working alternately as a scientist and engineer (or inventor).
Rather, as we will see, Bjerknes 's links to the world of technology came
more indirectly, through a career shift which brought hirn from theoretical
physics to meteorology, and thence to practical forecasting. As a student,
in the years 1889-1891, Bjerknes , studied mathematics and mathematical
physics in Paris and worked in the laboratory of Heinrich Hertz in Bonn .
He returned to his native Norway to complete his studies , and received his
Ph.D. there in 1892 for work on the application of mechanics to physics.
The following year Bjerknes took up a lectureship in mechanics at the
Stock holm Hogskola, where he remained untill907. He then accepted a
professorship in mechanics and mathematical physics at the Royal
Frederik University in Christiania (Oslo), where he remained untilI913 .
In the first phase of his career, Bjerknes pursued a research program
aimed at creating a mechanistic foundation for all of physics. More
specifically, he sought to unite electrodynamics and hydrodynamics, by
developing a mechanics of the ether. In this work, however, Bjerknes
found hirnself becoming isolated from the mainstream of interest in
physics. For whereas his research was rooted in a mechanical world view,
the dominant trend in physics was toward research programs rooted in
electromagnetic and energeticist world views.
In response to this situation, Bjerknes began to orient his research
toward meteorology. He had initially been introduced to problems in
meteorology by researchers in that field who saw relevance in his work on
192 EDA KRANAKIS
ether mechanics. And he found that he could have an impact and a degree
of recognition in meteorology that he could never hope to achieve in
theoretical physics, given the latter's changing focus. Bjerknes therefore
transformed his goal of developing a complete mechanics of the ether into
the goal of developing a precise mechanical physics of the atmosphere. In
1903 he formulated such a research program explicitly, and by 1906 it
became his principal, long-term research commitment.
On a professional level, Bjerknes's growing commitment to meteor-
ology was reflected in his acceptance, first, in 1913, of a position as
director of a newly established geophysics institute in Leipzig, and then,
in 1917, of a professorship in Bergen, Norway, attached to the
Geophysical Institute of the Bergen Museum. This commitment was also
reflected in his growing involvement with aeronautical and meteorologi-
cal associations. For example, he became a member of the Swedish Aero-
nautical Society, the Norwegian Ballooning Society, the Leipzig
Aeronautical Society, the Berlin Society for Aeronautics, and he became
closely involved with the International Commission for Scientific
Aeronautics, the major forum of the international meteorological and
aerological communities.
As Bjerknes shifted his interest to meteorology, he became increasingly
drawn into the realm of technology . For meteorology was at the inter-
section of a developing interorganizational network that linked scientists
interested in studying the physics of the atmosphere, like Bjerknes , with
military, business, and commercial groups whose activities required or
benefitted from more accurate weather forecasting : farmers, fishermen,
shippers, aviators, commercial airlines, etc. Representatives from many
of these groups participated actively in aeronautical and meteorological
organizations, such as the International Commission for Scientific
Aeronautics, and they lobbied (either through these organizations or
through the government) to have meteorological practice and research
defined around their needs.
World War I had the effect of considerably broadening and strength-
ening this interorganizational network. Perhaps most notably, weather
forecasting and the collection of weather data came to be seen as essential
for military planning, and for military and civilian aviation. In the words
of Bjerknes, WWI brought about "the militarization of meteorology" .37
A number of government weather bureaus, such as France 's, were placed
under military control during the war, and kept this organization even
after the war. In 1919, the British Meteorological Office, which had been
independent from the time of its organization in 1867, was placed under
the control of the Air Ministry and was given "for its primary and
immediate object the satisfaction of the requirements of aircraft" . 3 8
Bjerknes participated in extending and strengthening this inter-
organizational network. He served on national and international
committees concerned with Iinking meteorology to the needs of aviation.
These included a commission established in 1919 (und er the auspices of
the International Meteorological Committee) to investigate the
application of meteorology to aviation, and also the Norwegian
government's Commission on Aerial Transport. He wrote reports and
proposals for the military and for government outlining the advantages of
Iinking meteorology to military planning, farming, and the like. And
most importantly, he took steps to establish a practical weather
forecasting service that would meet the needs of these groups.
Bjerknes's decision to take up weather forecasting occurred shortly
after his return to Norway to take up the chair at the Bergen Museum. He
soon realized that the Bergen Geophysics Institute was not going to
receive the level of funding he had anticipated, and that his best chance to
obtain adequate resources to do research and train students was to orient
his work more closely toward immediate, practical forecasting needs. He
therefore lobbied in 1919 to obtain private and government funding to
establish a weather forecasting service at his institute. He had already
established such a service temporarily during the summer of 1918 with
funding from the Norwegian parliament. This temporary service was
intended mainly to provide weather forecasts for farmers, as a means to
help alleviate serious food shortages resulting from wartime trade
restrictions. In his 1919 proposal, Bjerknes pointed out the continuing
advantage to farmers of weather forecasts, and also the benefits it would
offer for commercial aviation, which was just being established in
Norway. (Two years earlier, in a secret report to the Norwegian minister
of defense, Bjerknes had also suggested the need to establish a field
weather service for the Norwegian military, and he suggested several ways
in which his geophysical institute could contribute to this effort.)
Bjerknes's proposal for a weather service was accepted, and thus he
established and became director of the West Norway Weather Bureau.
Since forecasting depended upon extensive data collection , Bjerknes and
his group became involved in establishing the necessary infrastructure for
the collection and transmission of weather data, e.g. in setting up obser-
vation stations and wireless telegraphy stations. Bjerknes also worked
194 EDA KRANAKIS
closely with farming and fishing organizations, shipping and aviation

interests, etc., so as to be able to structure his forecasting service around
their specific needs, which varied considerably. To give a few simplified
examples, farmers needed forewarning of sudden, daytime showers;
fishermen needed to know when and where it would be safe to fish; and
pilots needed to know weather conditions over a particular route several
hours in advance. Each kind of weather forecast required particular kinds
of observations and data.
Bjerknes's growing involvement with the "client network" developing
around meteorology, and his eventual immersion in practical forecasting,
motivated a transformation in his repertoire of practice. During the early
phase of his career, when he was still a mathematical physicist, Bjerknes
had a repertoire that included, among other things, a specific body of
mathematical and physical knowledge relating to ether mechanics, a
mechanistic world view, and a commitment to deductive, quantitative
mathematical theories (first of the ether and then of the atmosphere). The
transformation of this repertoire occurred in two principle stages. In the
first stage, Bjerknes brought his mathematical physics repertoire into
meteorology. His original research goal in meteorology was to use his
specialized knowledge of mechanics and hydrodynamics as the starting
point for developing a precise, quantitative mechanics of the atmosphere,
such that it would be possible to predict future atmospheric states.
Although this goal still rested upon a mechanistic world view and a
commitment to deductive, mathematical theories, there was one
important change: ether was an ideal fluid but atmospheric air was not.
Accordingly, to develop a precise mechanical physics of the latter
required large amounts of real data. Hence the first shift in Bjerknes's
scientific repertoire was toward a much greater concern for data
collection and the technological and organizational means to achieve this
goal. Theory, however, was to be the principal guide for data collection. 39
The second (and more profound) stage in the transformation of
Bjerknes's scientific repertoire occurred when he took up practical
forecasting. Although weather forecasting was linked to the discipline of
meteorology, it was in some sense a technological activity in that one of
its principal goals was to contribute to the operation and maintenance of
society's material infrastructurer" accurate weather forecasts were
supposed to keep ships from sinking, planes from crashing, etc . Bjerknes
took up forecasting with the idea of integrating it with his research
program in meteorological science. Yet in doing so he came to accept a
more practical, "technological" role for meteorological science. Hence-

forth his scientific research was carried out mainly in relation to the
immediate needs of forecasting. He still retained his stock of knowledge
relating to mechanics, but increasingly this came to serve as a qualitative
guide to physical thought rather than as a foundation for quantitative
theorizing. Meanwhile, other kinds of knowledge took on added import-
ance in his work - for example, knowledge about the resistance of fishing
vessels to wind, about aircraft and aviation, or detailed knowledge about
visible weather patterns in the sky at particular times and places. (Such
knowledge was largely irrelevant to Bjerknes's earlier research.)
Throughout this transformation Bjerknes 's commitment to scientific
theorizing remained intact, but his method of theorizing changed
radically. His earlier methodology had been predominantly mathematical
and deductive - that is, his theories were derived mathematically from the
laws of mechanics and hydrodynamics. Now his approach was more
pragmatic, empirical, induct ive, and qualitative. More specifically, his
theories and conceptions were now increasingly derived, on the one hand,
from the careful coIIection, coIIation, and analysis of weather data (which
in turn rested on cartographical techniques), and on the other hand, in
relation to the requirements and development of practical forecasting for
specific dient groups. His new repertoire also presupposed a different
relationship between theory and data collection . Whereas in his earlier
research, data coIIection was made dependant upon theor y, now theory
was made dependant upon data coIIection. And data coIIection, in turn ,
was now primarily structured around technological possibilities and
immediate forecasting needs. It should be noted that Bjerknes did not
entirely give up his old methods of theorizing . He still held on to the ideal
of making meteorology an exact science, and he still worked to integrate
concepts and models derived from forecasting practice into a more
precise, mechanics-based , mathematical theory. Yet his new methods
clearly had the predominant role in orienting the research program of his
group.
An example of Bjerknes's new approach to theorizing was the
elaboration by his group of the concept/theory of the polar front. On one
level, this new idea was the result of a change in their manner of
conceptualizing weather phenomena. Initially, the group, drawing on
Bjerknes's theoretical meteorology, had modeled weather phenomena in
terms of discontinuities in a two-dimensional windstream. Yet with their
growing experience in the practice of forecasting and in the mapping and
196 EDA KRANAKIS
analysis of weather data (which in turn were rooted in a particular

technica1-organizational system for data collection), the group began to
understand and model weather rather in terms of three-dimensional
surfaces of discontinuity - that is, in terms of the movements and inter-
actions of three-dimensional air masses distinguishable by a variety of
characteristics: pressure, temperature, direction of motion, etc. Initially,
they developed this conceptualization as a means to understand and
analyze cyclones. Subsequently they extended it to hemispheric pro-
portions, by postulating the existence of a polar front extending around
the upper reaches of the entire northern hemisphere. They theorized that
this front, in constant battle with masses of warm equatorial air, was the
principle determinant of weather patterns and cyclone formation in the
northern hemisphere.
On another level, the concept/theory of the polar front was also an
outgrowth of the emergence of a hemispheric transportation and
communications network. The concept of the polar front was first
enunciated at the end of 1919, after the first successful trans-atlantic
flights, and at a time when the international meteorologieal community,
in response to provisions laid down in the Paris Peace Treaty of 1919, was
working to structure an international, hemispheric network for the
exchange of weather data (extremely important for international air
traffic), and to develop techniques for long-term forecasts (of three or
four days). The polar front concept emerged and was developed partly in
relation to these needs and developments . Hemispheric systems of
aviation and communications both demanded and made possible a
hemispheric model of weather that in turn could be linked to a
hemispheric system of weather forecasting. And after enunciating the
polar front theory, the Bergen group immediately integrated it into a plan
for a circumpolar weather service, claiming that "polar front
meteorology" provided the key to accurate short and long-term forecasts .
CONCLUSlONS
The model of hybrid careers set out in this paper depicts them as being
linked, on one side, to interorganizational networks, and on the other
side, to the emergence of hybrid repertoires of practice. I illustrated this
model and its two organizing concepts by looking at the careers of Navier,
Thomson, and Bjerknes. Now we need to go a step further, however, and
use the evidence provided in the examples to reflect upon the

historiographical implications of the model. How, for example, does it
enhance our understanding of the science-technology relationship?
I believe that the model has four main benefits. First , through the
concept of hybrid repertoires of practice, it leads to a new and broader
perspective on the issue of what is transferred between the worlds of
science and technology. Most interpretations of the science-technology
relationship have focused only on a narrow range of possibilities. The
applied science model considered only knowledge transfers from science
to technology . More recently, the "two communities" model, while
accepting the existence of knowledge transfers, has primarily emphasized
the transfer of methods and organizational structures from science to
technology." In contrast, the concept of hybrid repertoires of practice
embodies the assumption that transfers between science and technology
simultaneously involve knowledge (data, concepts, theories, etc.),
research practices, and values (expressed, e.g., as priorities, commit-
ments, or as research ideologies).? Perhaps more importantly, however,
this concept leads us to recognize that the process of carrying out such
complex transfers demands a creative restructuring and reintegration of
the elements into a coherent whole .? Significantly (and as the term hybrid
suggests), this restructuring involves not merely the addition of elements
from one domain to an existing repertoire in the other domain, but also
the displacement of elements from the existing repertoire and sometimes
also the creation of new elements."
To illustrate, let us return to our examples. In all three cases the
emergence of a hybrid repertoire shaped the way in which the researcher
moved between theory, experiment, and technological practice (here
including forecasting). That is, it shaped the relationships among these
elements within the repertoire. Navier, by integrating his belief in the
supremacy of mathematics into his engineering repertoire, came to accept
a research ideology which saw mathematical analysis as the best guide to
technological practice, and which made experiment largely subsidiary to
theory . As a result, values and practices that were common among many
technologists of his day - such as the practice of problematizing tech-
nology in terms of design questions resolvable through testing or
experimentation - found no place in his repertoire. Thomson, by
integrating values and practices stemming from his technological
experiences into his scientific repertoire, came to the view that the
understanding of nature comes primarily through measurement and
198 EDA KRANAKIS
through technological knowledge and control. He accordingly came to

accept a research ideology which made mathematical theories of natural
phenomena dependant upon, and subsidiary to, both measurement and
technological applicability. As a result, Thomson rejected values and
practices that were common elements of the repertoires of many scientists
of his period. Perhaps most notably, he rejected the idea that mathema-
tical analysis could playa defining role in theory construction, the very
idea that underlay Maxwell's electromagnetic theory . Finally, Bjerknes,
by integrating a commitment to practical forecasting into his scientific
repertoire (that is, by accepting a technological role for meteorology),
also had to establish new relationships among the elements of his
repertoire. Scientific theorizing was now to depend not so much on
mathematical analysis as on data collection and on the practice and
immediate needs of forecasting .
It must be emphasized that in each of these instances, the creation of
a new, hybrid repertoire was an important stimulus to creativity. On the
one hand, to make the new repertoires coherent and successful required
innovation, and on the other hand, once established, they provided a basis
for innovative contributions to knowledge.:? Navier had to devise specific
techniques for using mathematical theories to derive useful technological
knowledge and design criteria. Thomson, in order to realize his vision of
precision measurement as a motor of scientific progress, transformed the
role, position, and importance of the scientific laboratory in teaching and
research. And Bjerknes, in order to integrate meteorological theorizing
with forecasting practice in a viable manner, introduced important new
techniques for collecting and analyzing weather data.
All of the three moreover made important contributions to knowledge
which were clearly rooted in their hybrid repertoires . Navier's theory of
suspension bridges, Thomson's reformulation of Newtonian mechanics,
and Bjerknes's polar front theory are just three examples among many.
These examples show that hybrid repertoires foster innovative knowledge
in part because they lead researchers to seek knowledge in relation to a
new set of criteria. Navier developed his theory of suspension bridges, not
just in relation to practical design requirements, but also in relation to the
research interests of French mathematicians and to his own vision
for theory-based technological knowledge. Thomson reformulated
Newtonian mechanics not only in relation to criteria originating within
the physics community, but also in relation to the needs and goals of
engineers. And finally, Bjerknes and his group developed the polar front
theory, not only in relation to criteria from mechanics and theoretical

meteorology, but also in relation to needs and criteria stemming from
forecasting practice.
The second advantage of the model of hybrid careers proposed here is
that it encourages us to think about hybrid careers not only individually,
but also collectively, as part of a broader social phenomenon involving
education, institutional development, and professionalization. More
specifically, this model recognizes that hybrid careers are rooted in
organizational structures and interorganizational networks that may
become more or less permanent. To the extent that this occurs, it leads to
the institutionalization, standardization, and social reproduction of
hybrid careers, so that they become no longer the exception but rather the
rule." All of the examples, if they do not explicitly demonstrate such a
process, at least point to it by implication. In the case of Navier, for
example, we saw that his career developed within an organizational
context in which years of formal, advanced mathematical and theoretical
training was a requirement for entry into the state engineering corps, and
in which the Academie des Seiences had an important role, on the one
hand, in determining the engineering curriculum, and on the other hand,
in judging the status and authority of contributions to engineering
knowledge . Needless to say, many engineers beside Navier also worked
within this system, and hence if we accept that it played an important role
in orienting his career, then we must ass urne that it similarly shaped the
careers of other engineers as well. "
The third advantage of the hybrid careers model is that it provides a
structure that allows us to analyze the interaction of science and techno-
logy as a historically contingent process involving social negotiation.
More specifically, the model shows that the links and boundaries between
the worlds of science and technology are negotiated at the organizational
level through the structuring of specific interorganizational networks , and
at the level of practice, through the development of specific hybrid
repertoires. The fact that these repertoires and networks are historically
contingent is made clear by the very different forms they take under
differing historical circumstances. The repertoires of Navier, Thomson,
and Bjerknes each established a distinct pattern for linking the worlds of
science and technology, one that fit with that individual 's circumstances,
experiences, and interests. Likewise, the interorganizational networks in
which the three worked also took different forms . Navier was part of a
network that linked a smalI, elite strat um of French technologists (those
200 EDA KRANAKIS
employed in the state engineering corps) to the French scientific

establishment. The structure of this network was such that Navier had
relatively little social or intellectual contact with technologists and
entrepreneurs working in the private sector . (If anything, such contacts
were avoided because those in the private sector were perceived to have a
distinctly lower social and intellectual status than the state engineers.)
Thomson, in contrast, was part of a network that specifically fostered
contacts between, on the one side, academic scientists and engineers, and
on the other side, technologists and industrialists working in the private
sector . Finally, Bjerknes was part of an interorganizational network that
made meteorological scientists semi-dependant upon dient groups within
government and industry, both through funding arrangements and
through the establishment of formal and informal relationships among
meteorological institutes, weather bureaus, and government agencies for
aviation and military affairs. In all of these cases, the character of the
networks at the local or national level depended upon many specific
historical conditions and processes: e.g. the French Revolution and the
debates over technical education which this transformation spurred; the
evolving social-industrial situation in Glasgow; or World War land its
effect both on Norwegian food supplies and on the financial position of the
Bergen Geophysical Institute.
The other main benefit of the hybrid careers model (and more generally
of the model of science and technology from which it derives) is that it
provides an explicit formulation of an important historiographical trend .
I refer specifically to the trend toward a more contextual form of historical
analysis in which science and technology are seen as social-cognitive
activities carried out within, and profoundly shaped by, specific historical
and organizational circumstances . The studies of Smith and Wise on
Thomson (Lord Kelvin) and of Friedman on Bjerknes are two excellent
examples of this trend, but others could be cited as well." In many of these
studies, however, the historiographical assumptions remain implicit: they
are used to structure the historical narratives, but are not explicitly
discussed or evaluated . The advantage of building these assumptions into
a formal model is that it forces us to confront and analyze them more
directly and it provides a framework in which to explore their conceptual
and theoretical implications. Doing so, in turn, can help us to see their
limitations and ultimately move beyond them .
University 0/ Amsterdam
NOTES
I Edwin Layton, 'Mirror-Image Twins: The Communities of Science and Technology in

19th-Century America', Technology and Culture 12(4) (October, 1971), p . 565. A model
similar to Layton's was discussed by Barry Barnes, 'The Science-Technology Relationship: A
Model and a Query' , Social Studies ofScience 12(1982), pp , 166-171.
, See, e.g. Walter G. Vincenti, What Engineers Know and How They Know It: Analytical
Studiesfrom AeronauticalHistory(Baltimore and London: Johns Hopkins University Press,
1990); Nathan Rosenberg and Walter G . Vincenti, The Britannia Bridge: The Generation and
Diffusion ofTechnological Knowledge(Cambridge, Mass. : MIT Press, 1978); Eda Kranakis,
'The French Connection: Giffard's Injector and the Nature of Heat' , Technology and
Culture 23(1) (January, 1982), pp . 3-38; Ronald Kline, 'Science and Engineering Theory in
the Invention and Development ofthe Induction Motor, 1880-1900' Technology and Culture
28(2) (April, 1987), pp. 283-313; Joan Bromberg, 'Engineering Knowledge in the Laser
Field,' Technology and Culture 27(4) (October, 1986), pp . 798-818.
J Although the "two cornmunities" model has generally been interpreted in such as way as to
emphasize the boundaries and cognitive differences between the worlds of science and
technology, Layton hirnself acknowledged that these two domains have become increasingly
intermeshed since the 19th century. He also hin ted at the need for a more integrated model of
science and technology: "In many modern social contexts physics and engineering have
become so intermixed that it is difficult if not impossible to sort them out into neat
pigeonholes . For example, in studying large, interdisciplinary research organizations, it is
often more helpful to think of physicists and engineers as part of a complex 'research system '.
Many men trained in physics 'do' technology, just as many men trained as engineers 'do'
science, including 'pure' or undirected research ." See Edwin T. Layton, 'Technology and
Science , or, Vive la Petite Difference' , Philosophy ofScience Association 11(1976), pp. 173-
183. The modell am proposing here is intended to allow for the possibility of the emergence
of this kind of integrated "research system",
4 In using the terms "science" and "technology" to refer to these realms of activity, I am in
effect adopting a convenient shorthand. I do not mean to suggest that the lauer are
historically unchanging.
5 My idea of science and technology as intersecting worlds owes a lot to the social worlds
concept that is central to symbolic interactionism. I was introduced to symbolic interactionist

thought and literature through the work of Joan Fujimura.1 would like to thank her for many
stimulating discussions. See Joan H . Fujirnura, 'Constructing "Do-able" Problems in
Cancer Research : Articulating Alignrnent", Social Studies ofScience 17(1987), pp . 257-293;
'The Molecular Biological Bandwagon in Cancer Research: Where Social Worlds Meet',
Social Problems 35(3) (June 1988), pp . 261-283; Bandwagons in Science: Doable Problems
and Transportable Packages as Factors in the Development of the Molecular Genetic
Bandwagon in Cancer Research (Ph.D. Diss., University of California, Berkeley, 1986).
6Layton, 'Mirror Image Twins,' pp. 577-578 .
7 The first example derives from my own research, while the other two are taken from the
excellent studies by Crosbie Smith and M. Norton Wise, on William Thomson, and by Robert
Mare Friedman on Vilhelm Bjerknes. See Smith and Wise, Energy and Empire: A
Biographical Study of Lord Kelvin (Cambridge: Cambridge University Press, 1989); and
Friedman, Appropriating the Weather: Vilhelm Bjerknes and the Construction ofa Modern
202 EDA KRANAKIS
Meteorology (Ithaca, New York : Cornell University Press, 1989).

• Archives Nationales, F I 42289, Service de Navier.
• C.-L.-M .-H. Navier to the Director General, Corps des Ponts et Chaussees, Chälons-sur-
Saöne, 21 June 1821, Archives Nationales, F 142289.
10 For an overview of Gauthey's career and work, see Emiland Gauthey, Traite de la
Construction des Ponts, C.-L.-M .-H . Navier (ed.) (Paris : Firmin Didot, 1809). This work
contains a biography of Gauthey written by Navier.
11 On the Societe Philomatique, see Maurice Crosland, The Society 0/ Arcueil (Cambridge,
Mass. : Harvard University Press, 1967), pp. 169-179.

12 I. Grattan-Guinness, Joseph Fourier, 1768-1830 (Boston: MIT Press, 1975), pp , ix-x ;
John Herivel, Joseph Fourier: The Man and the Physicist (Oxford: Clarendon Press, 1975),
pp . 128-129.
IJ A. Brunot and R. Coquand, Le Corps des Ponts et Chaussees (Paris: Editions du Cent re
National de la Recherche Scientifique, 1982), pp. 3-37; Charles Coulston Gillispie, Science
and Polity in France at the End 0/ the Old Regime (Princeton : Princeton University Press,
1980), pp. 479-498; Anne Querrien, 'Ecoles et Corps : Le Cas des Ponts et Chaussees, 1747-
1848,' unpublished MS, Bibliotheque de l'Ecole des Ponts et Chaussees; Coteile, Esquisse
Historique sur I'Institution des Ponts et Chaussees (Paris : Paul Dupont, 1849).
14 A. Fourcy, Histoire de i'Ecole Polytechnique, Jean Dhombres (ed.) (Paris : Belin, 1987),
passim ; Terry Shinn, Savoir Scientifique et Pouvoir Social: L 'Ecole Polytechnique, 1794-
1914 (Paris : Presses de la Fondation Nationale des Seiences Politiques, 1980), pp . 9-23 .
IS CharIes Dupin, Eloge de M. le Baron de Prony, Chambre des Pairs, 2 April 1840, pp. 22-
23,32-33; Herivel, Fourier, pp. 118, 129, 143.

I. Institut de France, Index Biographique des Membres et Correspondants de l'Academie des
Seiences (Paris: Institut de France, 1954). For a fuller discussion of French engineering in the
nineteenth century and its social and institutional context , see Eda Kranakis, ' Social
Determinants of Engineering Practice : A Comparative View of France and America in the
Nineteenth Century', Social Studies 0/ Science 19 (February, 1989), pp . 5-70.
17 Louis L. Bucciarelli, 'Poisson and the Mechanics ofElastic Surfaces,' in Michel Metivier,
Pierre Costabel , and Pierre Dugac (eds.), Simeon-Denis Poisson et la Science de son Temps
(Palaiseau : Ecole Polytechnique, 1981), pp . 95-104; Stephen P . Timoshenko, History 0/
I.
Strength 0/ Materials (1953, rpt. New York : Dover, 1983), pp. 119-122.
C.-L.-M.-H . Navier, Rapport aM . Becquey et Memoire sur les Ponts Suspendus (Paris :
Irnprimerie Royale, 1823), 2nd ed. Paris: Carilian-Goeury, 1830. Navier's theory of
suspension bridges is examined more fully in Eda Kranakis, 'Navier 's Theory of Suspension
Bridges,' in J .L. Berggren and B.R . Goldstein (eds.), ' From Ancient Omens to Statistical
Mechanics: Essays on the Exact Seiences Presented to Asger Aaboe' , Acta Historical
Scientiarum Naturalium et Medicinalium 39(1987), pp. 247-258 . The history of Navier's
Invalides suspension bridge over the Seine is reviewed in Eda Kranakis, 'The Affair of the
Invalides Bridge,' Jaarboek voor de Geschiedenis van Bedrijfen Techniek 4(1987), pp. 106-
130.
I' Claude-L. -M.-H . Navier to Sylvestre F. Lacroix, 9 October 1823, Lacroix MSS 2396,
Bibliotheque de l'Institut de France .
'0 Navier discussed these techniques in his Memoire sur les Ponts Suspendus, 2nd ed. p. 11.
11 This analysis of Kelvin's career is based on Smith's and Wise's comprehensive biography
and on two additional articles: M. Norton Wise and Crosbie Smith, 'Measurement, Work
and Indu stry in Lord Kelvin's Britain' , Historical Studies in the Physical and Biological
Seiences 17(1)(1986), pp . 147-173; and M. Norton Wise, 'Mediating Maehines', Science in
Context 2(1)(1988), pp . 77-113 . Specifie citations will be given only for direet quotations.
" Thomson's links with Walter Crum were partieularly close sinee he also married Crum's
daughter, Margaret, in 1852.
23 To understand the establishment of the Glasgow eha ir of engineering as part of a broader
transformation of engineering in Britain, see R.A . Buehanan, 'The Rise of Scientifie

Engineering in Britain,' British Journal for the History 01 Science 18(1985), pp. 218-233 .
24 An important souree for understanding Rankine's work is David F. ChanneIl, 'The
Harmony of Theory and Praetice: The Engineering Scienee of W.J .M. Rankine',
Technology and Cu/ture 23(1982), pp . 39-52.
" Quoted from Smith and Wise, p. 283.
26 Ibid. , p. 284.
27 These are Thomson's own words , quoted from Smith and Wise, 'Measurernent, Work,
and Industry in Lord Kelvin's Britain ', p. 150.

28 Smith and Wise, Energy and Empire, p. 491.
29 Ibid ., p . 19I.
3 0 Quoted from Smith and Wise, Energy and Empire, pp . 454-455 .
3 1 Thomson saw teehnological enterprise as the primary spur to the development of
preeision measurement. With respeet to eleetrieal measurements, for example, he remarked

that "resistance eoils and ohrns, and standard eondensers and microfarads had been for ten
years familiar to the eleetricians of the submarine-cable faetories and testing stations, before
anything that eould be ealled eleetric measurement had eome to be regularly praetised in
alm ost any of the scientifie laboratories of the world ." Quoted from Smith and Wise,
Energy and Empire, p. 455.
32 Smith and Wise, 'Work, Industry and Measurement in Lord Kelvin's Britain;' p, 167.
33 Smith and Wise, Energy and Empire, p, 130.
34 The idea oeeurred to Thomson du ring aperiod in which his brother James was writing to
hirn about methods for analyzing the effieiency and "mechanical effect" (i.e, work) of
steam engines and water wheels . Signifieantly, William Thomson used the same terrn,
"rnechanical effeet", in his analy sis ofthe eondueting sphere s. It was not a widely used term .
James Thomson took it up from Lewis Go rdon, who had proposed it as a translation of the
German, mechanische Wirkung.
J5 This point is further explained in Smith and Wise, 'Work, Measurement, and Industry in
Lord Kelvin's Britain,' pp . 158-159 , 167-169.

3 6 The analysis of Bjerknes's eareer is drawn from Friedman' s study, Appropriating the
Weather, and from Robert Mare Friedman, 'Constituting th e Polar Front, 1919-1920', Isis,
73(268)(September, 1982), pp. 343-363 . Specifie citations will only be given for direet
quotations,
3 1 Quoted from Friedman, Appropriating the Weather , p . 145.
38 Ibid.
39 Bjerknes's need for data was one ofthe rnajor factors that initially led hirn toward greater
involvement in practical meteorology.

40 The nature of the dient network surrounding rneteorology certainly supports such an
hypothesis.
41 It must be emphasized that many empirieal studies, even carried out with in the context of
204 EDA KRANAKIS
these models, provide evidence for a broader range of transfers between the domains of
science and technology .
.. By research ideologies, I mean normative ideas about how research ought to be done ,
what methods should be used, and what relationships should exist among research rnethods.
Research is here taken to include the full range of exploratory activities that scientists and
technologists engage in: theorizing, inventing, designing, experimenting , tinkering, etc.
Research ideologies offer guidance in answering questions such as the foHowing: is it best to
derive theories from empirical practice, or viceversa? How should theory and experiment be
linked in a research prograrn? How should new technological artifacts be designed - in
relation to empirical experience or to formal theories?
4J It should be noted that in the case of dual careers, like Navier's or Thomson's, the
transfer of knowledge, practices, and values between the worlds of science and technology
may result in not just one but two (interrelated) hybrid repertoires . For example, Thomson's
scientific and technological repertoires were not precisely identical, but because each shaped
the other, they did come to share many features. Due to space limitations , I did not examine
this phenomenon in the examples. Thus, in the case of Navier, I examined only his
engineering repertoire, while in the case of Thomson, I examined only his scientific
repertoire.
•• The existing repertoire may be either the individual's own, or it may be understood as a
shared repertoire, comprising knowledge, practices, and values common within a discipline
or field of expertise at a particular time and place.
• s This is not to suggest, however , that a hybrid repertoire would be a spur to creativity in
aH contexts or situations .
•• In asense, the professionalization of engineering can be seen as part of this process.
Moving the locus of technological training from the workplace to the university and
changing it from an apprenticeship into a formal education complete with books, lectures,
and instruction in experimental method s necessarily implies a broad transformation of the
technologist 's repertoire of practice.
47 In fact, as I have shown in another article, this was indeed the case. See 'Social
Determinants of Engineering Practice,' op. cit., Note 15.

•• Two other notable studies that particularly bear upon the question of the relationship
between science and technology are Stuart Blume, lnsight and Industry: On the Dynamics
0/ Technological Change in Medicine (Cambridge, Mass.: MIT Press, 1992); and George
Wise, Willis R. Whitney, General Electric, and the Origins 0/ V.S. lndustrial Research (New
York: Columbia University Press, 1985).
JOHN M . STAUDENMAlER . S.L
SCIENCE AND TECHNOLOGY : WHO GETS A SA Y?
In recent years historians of technology have increasingly agreed on one

major quest ion about the relationship of science and technology.
Technology's intellectual character cannot be reduced to "the application
of science," as if science created real knowledge and technology then
applied it to solve problems. Technological cognition, its own unique
form of knowledge, takes shape in a tension between generalizable
knowledge (called "theory" these days and "know-how" in an earlier era)
and the technical practitioner's capacity for pragmatic judgments. These
one-of-a-kind decisions are based on intimate knowledge of the
immediate situation (often called "skill" or simply "experience"),'
This suggests that longing for clearly-defined controlled environments,
for non political "sweet" projects, may be the engineers ' primary
seduction. When overcome with hankerings for Faustian bargains
engineering loses its balance and becomes brittle. Healthy technological
thinking, even at its most sophisticated "high-tec" levels, engages the
messy unpredictabilities of the situation at hand as an essential
component for creativity and does not flinch from them as unwelcome
interruptions in the design process.'
Put another way, technological thinking includes something more than
"rationality" or, as David Landes puts it, "the adaptation of means to
ends" ,.' It requires an intimacy with the societal and cultural context that
both transcends and anchors means-to-ends strategies. It must provide a
context of meaning within which means-to-ends strategies make sense,
wherein some strategies get accepted as worth the time and money. This
insight lies at the heart of a contextual approach to the history of
technology. One cannot und erstand any technology, so contextualism
argues, unless one studies how the design works (i.e ., its means-to-ends
rationalizations) and how its societal context explains why this particular
design came to be invested in at all." Increasingly, this contextualist
insight has become legitimate within the history of science where, again,
the motives and world view and vested interests of scientists have begun to
be respected as significant evidence for interpreting the resulting science
rat her than dismissed as peripheral exogenieties ."
205
P. Kroes and M. Bakker (eds.) , Technological Development and Science in the lndustrial
Age, 205-230 .
206 lOHN M . STAUDENMAlER , S.1.
It would seem, then, that the question of how the means-to-ends

rat ionalities of scientific theory and technological design relate to their
contexts has been pretty weIl settled. It may be, however, that the ghosts
of a David Landes vision of Western science and technology have not
vanished from the field. Enlightenment elitism that defines every sort of
thinking that cannot be reduced to means-to-ends strategy as
"superstition" has flourished for several centuries because, and this is the
point of interest here, it provides substantial benefits to scientific and
technological practitioners. One takes a short step indeed from the claim
that true science and technology must be allowed to operate free from
irrational or self-serving outside influences to the claim that scientific and
technological practitioners must be allowed to operate free from any
objection or critique by those outside their domains of expertise.6 It is not
by chance that Adam Smith's capitalism came to prominence just as
science and technology began to be treated as autodynamic enclaves of
pure rationality and beneficent "progress". Whether applied to the
market, to science, or to technology, Laissez Faire's iron law - "never
interrupt the working of the method by outside critique" - resides in
Western consciousness at the primordial level of symbol and rhetoric .
Despite the recent critiques noted above, it's seductive charms continue to
tempt usoSuch, at least, is the premise from which this paper proceeds.
In what follows I will sketch some parts of a complicated story, asking
how it came to be that such a chasm opened between those who design,
manage and decide the shape of society's dominant systems and the
ordinary citizen who has learned to conform to what has already been
settled in private. I will approach the question by calling attention to the
architecture of two factories, Jeremy Bentham's Panopticon and Henry
Ford's River Rouge. Both acquired the status of an archetypal symbol of
modernity and both reveal a central paradox for the modern West,
namely, that capitalism, for all its celebration of aggressive individual
creativity and initiative, turns out in practice to create systems whose
designs require passive conformity for the vast majority of those who
encounter them .'
FIO. 1I .-SECrIoH.
BUlLDI NG AND FURNITURE
INDUSTRY-I10USE ESTABLISHMENT,
FOIl 2000 I'RllSONS. OF A\,L AORS.
PA NOI'TI CON on CI;N Tl t A I••I NSI'ECTION PIUNCII'LE.

FIC. 11I.-G nolIND PLAN. Vl
o
[Tl
~ For th p r.'111:In:llirJIlI flf the lU""rr31 F i:;:u rr tl of t hi" Pur r.. see .. Out lint'
z
n
of .. w erk, ('ntitl,.,1 I · A I ~ I· F.R M AS Ao r .H E1'lT H lrROVr;o :", Upnth:am'. [Tl
" ' ork8, rnl. viii., 11. 3li9 to l'' 4:l9 . ;I>
Z
Th e Han~(". of IIrll..."I:l:rll ne.l C ,i l ~ :Ir,. r"'~PN' t i\', ·I,. 1lU tr roltf"rI t n r rrnm
o
)':nd to E nd of tll(' ,,,,Nnl W:alle. n ~ r ,l.il,itl'll in 11... f : N " I ' ~ l l I' I.I\s : tlll"J' e re -l
[Tl
here ffr rt'lt'nlr ,l In ('111 t hrnuJ.:h Il~- a I .int' II:tralll·1 In t1.f' ~i, 11' flf ll H' I'ul!~on : o
in the lkd ..gta~,.... ,,·1.31 ill rrl 'rt'1'l("nlr " :111 nHr" in thf' l . ra ll ~l.I . ill r rorolMl to :I:
be in IIt'O in Ihe U("8("riplion. Z
o
F!G. I . - E I,F:\ ' ATIO/'ö.
r"
o
Cl
0<
OB!
Fig. 1. Drawing of Bentham's Panopticon . From Panopticon; or, The lnspection-House: Containin g the Idea of a Ne w Principle
of Construction Applicable to Any Sort of Establishment, in Which Persons of Any Description Are to Be Kept Under
lnspection; and in Particular to Penitentiary Houses, Prisons, Poor-Ho uses, Lazarett os, Hou ses of Industry, Manufactories, 9
Hospitals, Work -Houses, Mad-Hou ses, and Schools by Jerem y Bentham, The Works of Jeremy Bentham ,
Val. IV (1838-43 edit ion ) facing page 39.
208 JOHN M . STAUDENMAlER, S.L
THE BENTHAM PANOPTICON: THE INVISIBLE WATCHER

AS THE NEW LOCUS OF POWER
Morals reformed - health preserved - industry

invigorated - instruction diffused - public burthens
lightened - Economy seated, as it were, upon a rock
- the gordian knot of the Poor-Laws not cut, but
untied - all by a simple idea in Architecture! - Thus
much I ventured to say on laying down the pen - and
thus much I should perhaps have said on taking it up,
if at that early period I had seen the whole of the way
before me.
(A new mode ofobtaining power 0/ mind o ver mind .
Jeremy Bentham")
Late in the 18th century Samuel and Jeremy Bentham designed the
"Panopticon", a "12 sided polygon formed in iron and sheathed in glass
in order to create the effect of 'universal transparency"'. 9 Samuel
Bentham invented the plan as an ideal factory for peasant workers in
Catherine the Great's Russia but it was his more prominent brother,
utilitarian philosopher Jeremy, who promoted the idea with the
entrepreneurial fervor of a missionary. He recognized that the model
could be applied not merely to factories, but to any institution aimed at
regulating the behavior ofthe unruly - orphans and the insane in asylums,
students in school, workers in factories, or prisoners. In 1787 he began a
several decade campaign for Parliamentary funding of a model prison, to
be managed by himself at a profit, in which a marginal ring of transparent
cells held isolated prisoners under the twenty-four-hour-a-day scrutiny
from the opaque inspectors' tower."
Bentham saw the prison as a factory ("a mill for grinding rogues
honest and idle men industrious"!') and scrutiny drove the mechanism for
manufacturing reform.
... the more constantly the persons to be inspected are under the eyes of the persons who
should inspect thern, the more perfectly will the purpose of the establishment have been
attained. Ideal perfection, if that were the object, would require that each person should
actually be in that predicament, during every instant of time. This being impossible, the next
thing to be wished for is, that, at every instant, seeing reason to believe as much, and not
being able to satisfy himself to the contrary, he should conceive himself to be so."
A radical imbalance of power divided invisible tower from isolated

prisoner. As Michel Foucault observes:
SCIENCE AND TECHNOLOGY 209
Bentham laid down the principle that power should be visible and unverifiable. Visible: the
inmate will constantly have before his eyes the tall outline of the central tower from which
he is spied upon . Unverifiable: the inmate must never know whether he is being looked at at
any one moment ; but he must be sure that he may always be so ... in the peripherie ring, one
is totally seen, without ever seeing; in the central tower, one sees everything without ever
being seen. 13
Onee imbued with the self image of living under the serutiny of unseen
eyes, twenty four hours a day, the prisoner eould be released into society.
Bentham intuitively grasped the nature of a profound revolution in the
West's understanding of how the individual relates to the governing
struetures of society. His Panopticon plan abandons the earlier judicial
foeus on the eriminal aet (with its appropriate punishment) and
eoneentrates on the criminal (and his/her need to be eorreeted) . In the
older way of doing things, if I am judged guilty of a erime I beeome liable
to a specific punishment, but I am not defined as a defeetive human nor
am I excluded from the larger human eommunity in principle. As in the
Greek term for sin, hamartia ("missing the mark"), my guilt has to do
with the aet I eommitted. Onee my debt is paid, it is assumed that I have
a plaee within the human community. This even holds in eases of eapital
punishment for soeieties that provided explicit reeoneiliation prior to
exeeution. Onee absolved, the repentant sinner died reunited with the
eommunity of the Chureh which, nevertheless, exaeted the death penalty
that aeerued to the eriminal aet. Presumably, both exeeutioner and the
exeeuted would hold eommon status in the heavenly eommunity in which
both believed and on which the praetice was based.':'
Bentham's plan abandons the assumption that erimes ean be paid for
by punishment without eroding the ordinary humanity of the eriminal.
The Panopticon redefines eriminals as defeetive and needing eorreetion.
Foueault sees this shift from punishable aet to defeetive eharaeter
pervading nineteenth and twentieth eentury European penal praetice;
judges have "taken to judging something other than erimes, namely, the
'soul' of the eriminal". 15 Judging the interior of the eriminal and applying
eorreetive remedies implies some exterior standard against which one is
measured and that in turn implies some person or group who get to set the
standard.
The unstudied rhetorie of those who claim the elite status of therapist
frequently reveals the depth of eontempt for those deemed in need of
repair that is implicitly part of this social arrangement. Thus Karl
Menninger writes:
210 lOHN M . STAUDENMAlER, S .1.
We, the agents of society , must move to end the game of tit-for-tat and blow-for-blow in
which the offender has foolishly and futilely engaged hirnself and us. We are not driven, as
he is, to wild and impulsive actions, With knowledge comes power, and with power there is
no need for the frightened vengeance of the old penaology. In its place should go a quiet,
dignified, therapeutic program for the rehabilitation of the disorganized one, if possible, the
protection of society during the treatment period, and his guided return to useful citizenship,
as soon as this can be effected. "
One could ask, of course, how any group of human beings acquires the
right to set standards for the correction of the character of other human
beings. How do some human beings become "the" agents of society and
members of this imperial "we", while others remain locked in the third
person (i.e., "the disorganized one" who needs "our" guidance to be
returned to useful citizenship)? The device ofthe invisible tower preempts
such questions by masking the power relationships at work here. The
desideratum of Benthamite incarceration is not a community wherein
identifiable people negotiate the norms of acceptable human behavior.
Instead, the Panopticon inculcates the experience of invisible scrutinizers
who, indeed, need not be in the tower at all.
Echoes 0/ Bentham: Computerized Scrutiny and Computerized Tests
Bentham bitterly regretted Parliament's refusal to implement his insight.

Nevertheless, his basic concept has flourished in Western societies ever
since, most vividly, perhaps , in George Orwell's 1984. Orwell mirrors the
Panopticon with the omnipresent telescreens. His dystopian dictatorship,
Oceania, scrutinized all citizens all the time to prevent those forms of
mysterious and unpredictable freedom which can surprise and confound
the state's planning. As in Bentham's prison, of course, one can never be
sure whether "Big Brother is always watching" at any particular moment.
This same desire for omnipresent scrutiny as a form of control has
been documented by Harley Shaiken and Shoshanah Zuboff. Shaiken has
articulated the concept of "over computerization" by which he means the
use of computers to increase managerial control at the cost of decreasing
productivity and efficiency. Zuboff documents repeated instances of
managerial loss of nerve in the face of radically expanded access to
computer data bases by all levels of the corporate hierarchy . Despite
significant increases in productivity and efficiency that follow such open
access policies, democratization of information access is often aborted by
acts of autocratic revisionism on the part of upper management."

On a more mundane level, an even more significant instance of the
pattern has emerged; the immense popularity of standardized, computer-
scored tests (e.g., SAT or GRE examinations) underscores the pervasive
character of the cultural change implicit in Bentham's rejection of the
punitive in favor of the reforming prison.
Beginning in the 19th century, written and numerically scored tests
began to be favored over the "performance" test. Performance evalu-
ations (e.g., medieval disputations, apprenticeships, oral examinations
generally) require interaction between judge and judged; their interactive
structure presumes that those on both sides of the evaluati ve divide share
a common humanity. I pass a performance test when some judge (or
judges) include me, by approving my performance, within the community
which they have been authorized to represent. The inherent subjectivity of
such performance testing - seed bed over centuries of various abuses such
as old-boy networks and racist double standards - is moderated
somewhat by the fact that I can physically touch and interact with those
who claim the authority to judge me. Numerically scored tests, in
particular when they took the true-false or multiple choice form early in
this century, radically erode the communality and sensuality of the
relationship. When I pass or fail a selection test I neither see or meet, nor
can I influence, my judges . Like the invisible scrutinizer in Bentham's
tower, those who wrote the standardized questions, among whose pre-
designed responses I select, remain anonyrnous."
This paper 's brevity does not allow a full exploration of the bipolar
roles created by the Panopticon. It remains important, however, to at
least sketch out some of the piaces where it appears in the 19th and 20th
centuries lest we underestimate its extraordinary hold on Western
consciousness. All of us, when in our roles as ordinary citizens, find
ourselves alternating between the role of the invisible wateher at the center
of power and the marginalized and isolated object in the peripheral cell.
Besides our experience as participants in pre-designed, standardized tests
noted above, we might consider a handful of examples where the same
pattern is manifest in the United States.
212 JOHN M. STAUDENMAlER, S .1.
Etiquette, the Theatre and Electronic Media: From Interactive to

Passive Audience
Beginning about 1830, etiquette books began their remarkable rise to

popularity in the United States. They taught those aspiring to the middle
class how to avoid misbehaving in public, a skill that Ervin Goffman and
lohn Kasson call "impression management". The manuals taught, for
example, nose decorum. You don't blow your nose in public if at all
possible; you don't fondie your nose, and you certainly don't pick your
nose and wipe your hand on your trousers. The books proscribed staring,
shouting, singing, or humming out loud. Civilized adults kept unruly
passions - sexual feelings, anger, even grief and joy - locked inside. One
neither wept nor exulted in the streets."
Earlier village life - for all its gossip and the long, sometimes un-
forgiving, memories that such an enclosed society fostered - rooted public
identity in storytelling. People were known, not on the basis of the
moment's immediate behavior, but rather according to cumulative stories
that recalled a lifetime of events. Impression management, on the other
hand, rests on the assumption that I walk the city as astranger subject to
the scrutiny of other strangers. Thus etiquette literat ure redistributes the
Panopticon roles. The anonymous scrutinizers have left the central tower
and inhabit the streets of the city around me. From one angle, I am the
marginalized and isolated outsider who is anonymously judged. From
another, I manage my exterior impressions from the invisible tower of my
hidden inner self.
A dramatic change in acceptable public behavior after 1840widens the
gap between wateher and watched in the context of public performance.
Imagine, for example, going to the music hall or theatre about 1840 in
New York. Audiences, raucous throngs that included all social classes in
the same building , routinely interrupted the musicians by cheering,
shouting and pounding the floor with their feet and sometimes demanding
repetition of an especially-liked passage of the music. Patrons booed ,
hissed, and threw rotten fruit or eggs at the play's villain. Lawrence W.
Levine offers the following description of audience reaction when a
visiting Italian opera troupe cut the final scene of an 1837 performance of
Rossini's Semiramide without prior announcement. The New Orleans
Picayune described the outcome of management's attempt to stop the
wild uproar by darkening the hall and driving the audience out.
t'was the signal for the demolit ion of everything they eould lay their hands on .... The
drapery around the boxes was torn , the eushions in the pit ripped open , the seats broken, and
ehairs were flying in all direet ions .
The next night, the chastened company performed "the last note that ever
Rossini cornposed" . 20 Audience-performer interaction was the norm and
public performance sometimes approached the chaotic.
After the Civil War, however, a new discipline began to take hold .
Thus, noted conductor Theodore Thomas insisted on silent audiences,
sometimes turning to stare them into submission before continuing the
performance. He and others like hirn conducted stern lessons in
conformity and passivity. This increasingly powerful movement of
audience reform reverses the polarities of Bentham's prison. The watched
hold the position of power at center stage. Losing the power to interact,
the audience is marginalized in the act of watching."
The new style of audience conformity started to take hold just as
electronic media began to revolutionize the very structure of public
discourse. Before Samuel Morse's invention of a telegraphic code, and for
the most part before the organization of the national wire service about
1870, information rarely traveled faster than a horse could trot. "The
news" reported on a minuscule universe, one's village and a surrounding
countryside perhaps thirty miles across. Messages from beyond horse-
range arrived long after things happened and their rhetorical form
differed accordingly. Pre-electronic newspapers published "corre-
spondences" , leisurely essays for readers who needed subtle details to
understand the gradual unfolding of far away events. On the local scene,
however, news writers and news readers shared the same living space.
Thus, the reader could ordinarily supply a host of nuances that the printed
account only suggested and disagreements about interpretation could be
settled right in town. Like the boisterous theater and music hall audiences,
newspaper readers could actively intervene in public discourse.
Beginning about 1870 national telegraphic wire services began to
change all that. 22 In 1876, for example, if I lived in Chicago I would
probably have read in the morning paper about the "Molly Maguire"
trials hundreds of miles away in the eastern Pennsylvania coal region. I
would learn that "the Mollies" were anarchistic Irish miners conspiring to
violently destroy the lives and property of coal mine owners. I would have
read, the morning after it happened, that twenty four were convicted and
ten hanged. I would not have known, however, that most historians have
since come to interpret the trial as a frame-up and that the key witness was
a Pinkerton detective in the employ of the owners. 13 The wire service news
crafters did not choose to include that perspective and I, at my Chicago
break fast table , lived too far off to know more than the wire service told
me. Live radio news beginning in the 1920s and more recent television
coverage took instantaneous news one step farther. Instead of reading
about news one day late, the audience could now "participate" in events
as they happened. Despite this dramatic intensification of audience
involvement, however, the listeners or viewers gained not one whit of
active power to shape public discourse .
Arecent exception demonstrates the rule. During the Reagan-Bush
administrations, while the New York Times and other mainstream media
accepted the administration interpretation of Central America (e.g.,
struggling democracies in Salvador and Guatemala vs vile dictatorship in
Nicaragua), a grass roots network communicated a distinctly different
reading of events and significantly influenced national policy as a result.
One has only to note, however, the enormous individual, group, and
church efforts involved to recognize the untypical character of the
example. Whether I favor or oppose the mainstream ideology is not at
issue here. Whatever my ideology vis avis the electronic media version at
any given time, I typically relate to electronic discourse as a passive,
isolated and powerless member of the audience , the very model of a
Benthamite prisoner in reverse, frozen in the act of watching. Even when
I shout at the TV because of a particularly odious ad or newscaster
remark, nothing public or civic happens . It wears me out to try to imagine
that I might change things.:"
Advertising and the Programmed Self
Twentieth century advertising, the most important single offspring of

mass media, demands special attention. The first three decades of this
century saw a revolution in advertising style that originated in, reinforced ,
and eventually restructured the patterns we have been considering .
Nineteenth century advertisements tended to take the form of rational
dialogue that assumed a basic equality between advertiser and reader .
With the noteworthy exceptions of patent medicine, and P .T. Barnum-
style fantastic entertainments, sales were thought to result from a
description of product qualities or the simple announcement of available
merchandise. By the early twentieth century, however, gradual changes in

strategy began to coalesce in a new style. Just as Foucault's modern penal
system shifted from assigning guilt and penalty for criminal acts to
judging the inner soul of the criminal, so the new advertising turned away
from the product to focus on the consumer and the benefits bestowed by
the product. Emotion began to replace reasoned argument as the
preferred rhetorical style; the targeted consumer was presumed to be
irrational and inept. Advertising historian Roland Marehand describes
the basic mentality.
In viewing the urban masses, advertiser s associated consumer lethargy as much with weak-
kneed conformity as with cultural backwardness.... Emotional appeals succeeded because
only by seeking this lowest common human denominator could the ad vertiser shake the
masses from their lethargy without taxing their Iimited intelligence."
Consumerist advertisers go Bentham's invisible watchers one better; they

search out the hidden motivations of "the consumer" to program them
into conformity with the requirements of some unannounced economic
interest. Insofar as consumerist advertising succeeds, itself a matter of
considerable debate, I become a marginalized outsider to my own inner
self. 26 The intent of the advertiser's shift from rational inter-subjective
dialogue and announcement toward emotional manipulation is precisely
to disconnect the consumer from his or her subjectivity ; the inner self-
with its tangle of integrity, passion, violence, and nobility-remains
inaccessible even as it is manipulated by the advertisements.
FORD 'S RIVER ROUGE : STANDARDIZATION WITHIN

A GUARDED PERIMETER
He said we would have to have a model [of the

proposed Ford Museum building) made, so we had a
model made and it showed the balconies, naturally,
and the basement , and he said, "What is this up
here?" I said, "That is a balcon y for exhibit. He said,
"I wouldn 't have that; there would be people up
there, I could come in and they wouldn 't be working ,
I wouldn't have it," He said, " I have to see
everybody ," Then he said , "What's this? " I said,
"That is the basement down there, which is necessary
to maintain these exhibits and to keep things which
216 JOHN M . STAUDENMAlER, S .1.
you want to rotate, etc." He said, "I wouldn't have

that; / couldn't see those men down there when /
came in. You have to do the whole thing over again
and pur it on one floor with no balconies and no
basements ." I said, " Okay" ."
Across the Atlantic and more than a century after Bentham, a second
factory design came to be seen as an icon of the spirit of modernity. This
factory, however, acquired the capital investment and physical reality
that eluded Bentham. Within a few years of its completion in 1923, the
Ford Motor Company's massive River Rouge plant achieved mythic
stature and world acclaim as the ultimate expression of "Fordismus", the
triumph of rational efficiency over nature and the burdens of life.
Tourists from around the world, by the hundreds of thousands made
pilgrimage to the Rouge in the late Twenties. German engineer, Otto
Moog was not unusual when he recorded his impressions in language that
combines an almost schizoid mixture of quasi-religious intimidation and
awe with an exultant sense of liberation.
No symphony , no Eroica, compares in depth, eontent , and power to the music that
threatened and hammered away at us as we wandered through Ford's workplaces ,
wanderers overwhelmed by a da ring expression of the human spirit. ,.
Solipsism and Power: Ford's Obsession with Control
When construction began along the banks of the Rouge River in 1915,
"Henry Ford" was already a household name . More than half the autos
in the United States were Model Ts, arguably the most successful match
between a single technical design and its societal context in recorded
history . For weIlover a decade, Ford sold the ugly, durable vehicles as fast
as they could be manufactured. Ford had already begun to capture world
attention when, in 1914, his simultaneous completion of the moving
assembly line and doubling of wages ("The Five Dollar Day") stunned
competitors and intensified his image as industrial genius and working
man's friend .
Less visible was his fixation on contro!. Workers, for example, became
eligible for the $5 Dollar Day labor reform package only when they
submitted to and passed inspection about intimate details of their
personal lives. Members of the newly created Sociological Department
'~r, "'" ;''-' . .clj\.lI1_ aJ.• •
~,... 1
",,,,...1 .....:, 1oCl.1XlO _
.. 1_ ~) "U ..
..s.',.lIJ.ar.' ,,"I'l '.Lw.I.
cn
o
rn
Z
o
rrr
;l>
Z
tl
o-j
r:n
o
::r::
z
o
r-
o
o
><
Fig. 2. The Highland Park Plant opening on to Woodward Avenue . From the collections of Henry Ford Museum & Greenfield
Village (Neg. no . 0.3139) . Reproduced with permission.
IV
.....
-.l
218 lOHN M . STAUDENMAlER. S.1.
checked cleanliness, debt, drinking habits, etc. A failing grade meant that
the profit sharing bonus was put in escrow until the worker mended his
ways. Failure to comply eventually meant firing." Seeking to escape
dependence on outside suppliers, Ford used the Rouge river's access to the
Great Lakes and adjacent intersecting rail lines to create the logistical
heart of a mine-mouth-to-dealership empire. He purchased mines and
forests, a rubber plantation in Brazil, a rail line and a Great Lakes
shipping line so that by the mid 1920's Ford boats and trains carried Ford
iron ore, hardwood and other inputs to the largest industrial plant the
world had ever seen.
lronically, while the Rouge appeared to the world as the symbolic
capstone ofFord's world-class technological triumph, the design concepts
on which it was based reveal unmistakable signs of Ford's retreat into a
solipsistic world that excluded of those whose independent thinking
threatened hirn. Thus, in 1919, after major stockholders sued (and won)
because of Ford's practice of diverting potential dividends back into
company expansion, Ford conducted an elaborate and deceptive strategy
for buying them out. Almost simultaneously, three of "his ablest
lieutenants" (C. Harold Wills, John R. Lee, Norval Hawkins) resigned
under pressure. 30
During the twenties, Ford's reclusive tendencies deepened. Fairlane ,
the Ford mansion completed in 1916, stood on the banks of the Rouge
river in Dearborn miles from most of Detroit's elite society who lived
across town along Grosse Po inte's mansion row. After his national
humiliation during the 1919Chicago Tribune trial , he withdrew more and
more into the company which he now totally controlled and, as the decade
continued, to his personal playground, the Henry Ford Museum and
Greenfield Village. For some years after it became operational, the
Museum and the Village were not open to the public . The complex served
as a small trade-based boarding school. Private groups and individuals
were sometimes permitted to visit the collections but no policy of public
admission was implemented until June 22, 1933.3 1
One ofthe Rouge's design departures from its predecessor at Highland
Park would later become a world famous symbol of Ford's rejection of
debate and dissent. Just as Ford pursued integrated control of inputs
through his network of transportation lines converging on the Rouge, so
he sought even more control over workers than the house-to-house
inspections and in-factory spy networks of the 1914 labor reforms
provided. The earlier Highland Park plant opened directly onto
Woodward Avenue, leaving Ford management with no say about who

mingled with workers on the public streets fronting the factory. Plant
accessibility meant unwanted interference. In dramatic contrast, the new
Rouge compound was insulated by a fully fenced perimeter and tightly
guarded gates . The Miller Road Gate - 4 became world infamous when
photographers caught Ford thugs beating VA W activists Walter Reuther
and Richard Frankenstein in 1937's "battle of the overpass".
Independently minded workers seeking to organize a union were
anathema to Ford; they apotheosized the intrusion of pluralism into the
ideologically standardized interior of the Ford universe.
Fig. 3. The Battle of the Overpass . From the collections of Henr y Ford Museum
& Greenfield Village (Neg. no. 0.4951). Reproduced with permission .
Ford's Aesthetic: The Clean Machine
Ford's avers ion to independent thinking is echoed in his habits of

displaying industrial technology. His great love affair with technological
symbols during the twenties - the multi million dollar Henry Ford
220 lOHN M. STAUDENMAlER, S .1.
Museum and Greenfield Village - revealed a powerful strain of romantic

nostalgia and an equally powerful commitment to contemporary
technocratic motifs . Greenfield Village ignored twentieth century
technologies almost completely and, indeed, made expensive hash of any
historical chronology. Ford bought what he liked and installed it. Shrines
to American heroes dotted the landscape: Abraham Lincoln's early court
house, hornes of Noah Webster, Ford hirnself, and the Wright brothers,
and the jewel of the collection, a worshipful reincarnation of the
laboratory complex at Menlo Park where Thomas Edison had invented
the electric light a half century before. 32
The adjacent Ford museum aimed more at aesthetically pleasing arrays
of artifacts (most notably: steam engines, automobiles, locomotives,
agricultural equipment, machine tools, and domestic appliances), each
series arranged in chronologically ascending order to demonstrate the
march of inventive progress . In both places, however, as weIlas in almost
every other part of his domain, Ford went to considerable lengths to
display large scale machines as sensuous, almost numinous, icons. In the
Village, he ordered the walls opening into the steam engine rooms of the
Loringer gristmill and adjacent Armington and Sims machine shop
changed from the original wood to glass so that he could watch the
machines work as he drove by. The powerhouse at the Highland Park
plant was deliberately displayed to Woodward Avenue passers by through
walls of showroom glass; inside, the dynamos were enthroned amid
gleaming brass and tile and not a few observers of Ford's secluded
"Fairlane" estate have noted that its only really elegant building was the
hydro-power plant, that displayed its dynamos in similar fashion."
Ford's industrial aesthetic, so like the design of the Rouge plant in its
avoidance of the messy political side of technology, can be understood as
the eccentric aberrations of a man whose rise to farne and fortune
permitted hirn to indulge in a growing obsession with privacy and personal
contro!.
From another perspective, however, Ford's ambivalent mix of
liberation and repression, so aptly captured by Otto Moog and similarly
articulate visitors to the Rouge, fits quite closely other idealizations of
progress in the twenties and thirties. Perhaps the most striking symbolic
representation of popular ambivalence about technological triumphs
would appear in the main foyer of the Hall of Science at the 1934Chicago
"Century of Progress" Exposition, itself an astonishing romanticization
of science and technology during the depth of the Depression . Visitors to
Fig. 4. The Dynamo Showroom from the Highland Park Plant. From the collections of
Henry Ford Museum & Greenfield Village (Neg. no. 833.72). Reproduced with permission.
the Fair's Hall of Science were met in the foyer by the "Fountain of
Science" with Louise Lentz Woodruff's three-piece sculpture, "Science
Advancing Mankind" . Two life-sized figures, male and female, faced
forward with arms uplifted. Both were dwarfed by the massive figure of
a metallic robot twice their size. In the words of Lenox Lohr, general
manager of the exposition, the robot typified "the exactitude, force and
onward movement of science, with its hands at the backs of the figures of
a man and a woman, urging them on to the fuller Iife." The sculpture's
iconographic ideology was reinforced by the official Guidebook's
stunning, bold-faced thematic motto: SCIENCE FINDS, INDUSTRY APPLIES,
MAN CONFORMS. 34
The Chicago Fair's technocratic ideal is rooted in a violent disjunction
between the combined force of "Science" and "Industry" (itself a
conflation of business and technology) on the one side and "Man" on the
other. The role of "Man" in the modernist equation is not to "Critique"
fallible decisions made by scientists, business managers or engineers; it is
not even to "Use" the new technologies. The Role of "Man" is to

"Conform":"
SYSTEM VS SENSUALlTY : SCIENCE AND TECHNOLOGY

AS CUL TURAL FORMS
All these stories abound with instances of the same paradox: Ford's
Model T provides geographical mobility for an entire generation of
people with modest means, giving them access to sophisticated
transportation technology, both cheap and easy to repair. Meanwhile the
much more sophisticated system that produces them encloses Ford
workers in prison factories hedged about with enforcers and impermeable
fences. Bentham's Panopticon would free the criminal from torture and
dungeon but offers re-instatement into society at the terrible cost of
programmed conformity to a mythic, omnipresent, and invisible
scrutinizer. Twentieth century citizens break free from the suffocation of
minuscule village perspectives through the mediation of global and
instantaneous electronic media. At the same time their capacity for active
participation in public discourse and the political order erodes.
It is a telling irony, one that reveals a great deal of the strengths and
liabilities of Western-style modernity, that both scientific and
technological practice themselves often mirror Bentham's tower and
Ford 's policed perimeter. Panopticonism and Fordism work to interdict
negotiation about system design and operation between scrutinizer and
prisoner, management and labor, the governing elite and the governed, in
short, they shield experts from the intervention of non-elites. In precisely
the same way, the scientific method's controlled variable environment
with its canon of replicability excludes non-experts from the workings of
the method. Science has its own aesthetic, a sanitized place cleansed of the
untrustworthy presences of bias, vested interest, emotion or tradition.
The controlled variable experiment, therefore, constitutes a revolutionary
new form of cultural space: an inaccessible interiority cuts off the working
of the experts' method from the turbulence of outside society.
This protected environment has, as we have seen, developed into a
multi-form and pervasive cultural force. Its most important influence
may weil be the radically increased social distance between decision
centers, where options are debated, taken, and eventually legitimated,
and the periphery, where non-decision-makers endure the results of the
hidden process. That societal chasm is filled by a constructed mechanism,

sometimes called "the system", which demands conformity as a take-it-
or-leave-it price for participation. Sometimes the distance between
margin and center can be maintained with a smiling face, as with the
entertaining demeanor oftelevision personalities, or with no face at all, as
in anonymous standardized tests. At other times, as in Ford's closely
guarded Rouge plant, the distance can only be maintained by brute
physical force. In all cases, the primary benefit acquired by distancing the
decision center from intruders remains the same. Some insider elite,
scientific, technological , economic or political, is given the space within
which to pursue the detailed and arduous tasks of system design."
Of course, the act of designing requires a protected space. At some
point debate must stop and decisions crystalized if the creative process is
to come to term . Nevertheless, protected design space is a mixed blessing
in the strict sense of the term "mixed" and it is the business of this paper
to call attention to what is gained and lost by the sociallegitimacy that has
accrued to it in the past two centuries.
The Good News . Western science and technology have eonverged as

powerful influences to foster an increasingly global consciousness that
transcends the limitations resident in any local scene. Scientific
communities, for example, have developed a eosmopolitan tradition of
multilingual cooperation, in part due to their insistence that valid
conclusions must be open to replication by peers who live somewhere else.
Transportation systems (air, overland, and water) have achieved such
high levelsof reliability and speed that they now constitute an interlocking
network for moving people and freight virtually anywhere in the world.
Electronic media, of course, move information - text, sound , and image
- more rapidly still so that the entire set of network technologies work in
coneert to break human transactions free from loeal constraints."
By their commitment to rationality and speed, and eonsequent
alignment against the local, the anecdotal , and the sensual, Western
scienee and technology tend to dampen the periodic swings toward
fanaticism which flourish in the humid soil of enclosed, sensual and cultic
societies. It is an achievement of great importance. James Jones'
"Jonestown" has eome to serve as a tragic ieon in this regard . That half
a thousand people could be induced to commit communal suicide surely
had something to do with their jungle isolation. Outsiders had no
opportunity to critique those in the eommune and shake them from their
224 JOHN M. STAUDENMAlER, s .r.
mesmerized fanaticism. What was lacking, in scientific terms, was the
application of the canon of replicability . Outside observers should have
been given voice to contest the claims of James Jones.
The Bad News. Disempowering the local and indigenous, however, can
be powerfully seductive for elite design constituencies. Note that the
patterns of culture so briefly sketched above reveal a cultural context
wherein small elite power centers pair off against large arrays of
marginalized conformers. Whether in the Bentham panopticon, the
increasingly class-divided theatres, the twentieth century electronic media
audiences, or a Ford factory, these systems of cognition and discourse all
insure what Bentham intuitively insisted on: that those on the margin not
be able to see or actively influence those at the command central tower.
This tends to play itself out in Science with contempt for local
traditions of wisdom, whether pre-Industrial or non-Western, with the
conviction that non-experts have nothing to teach the design elite in
question . Managers of technological systems, for their part, frequently
invest in designs aimed at preventing workers and "the public" from
having access to managerial decision making processes. In other words,
elitist control sometimes masquerades as a technical requirement when it
merely enforces a hankering for working space that has been cleared of
outsider critique."
So we come to the question; do science and technology as ordinarily
practiced constitute a liberating force or an imprisoning one? Insofar as
scientific and technical elites find ways to render their value-Iaden
judgments about priorities accessible to outside critique and pluralistic
debate, their work will flourish as an essentially beneficent influence
within society at large. lnsofar, on the other hand, as such elites succumb
to the seduction of the Bentham Tower and the Ford Fence, they
ironically replicate the very evils they claim to overcome. Inside the tower
and behind the fence experts easily succumb to the fleshy temptations
inherent in any enclosed local environment - old boy networks and the
blind biases of unchallenged assumptions. The structure of the
protecting claim to dispassionate and value-free objectivity and the
rejection of human pluralism implied therein can easily mask the
question of power. 39
To summarize in necessarily over simple terms: the policing of the
method, once the value 01 its project has been agreed on, needs to be the
establishes the value of one design over another must include non-elite
outsiders.
This is very hard work and I do not intend to minimize its difficulty.
Indeed, the primary purpose of this paper is to call attention to the
cultural patterns which render it so difficult for Westerners to find the
patience and the nerve to engage in tough debates about scientific and
technical priorities, about the allocation of the increasingly scarce
resources available to our societies, and about the determination of the
questions: "Who wins, who loses, and how much does it matter?" The
cultural trends toward a Benthamized public conformity-where
individuals shrink from the responsibilities of adult citizenship in the
public arena-run deep . They are embedded in our affectivity and our self-
understanding and they shape our expectations about how much can be
hoped for from the public order. Citizens of Western societies must
reopen Bentham's tower and break through Ford's fences so that we can
risk a rebirth of the turbulent and unpredictable forms of genuine civil
discourse as the millennium in which Western culture achieved global
power draws to a close. Lacking the courage to do so would very likely
insure that the next millennium will not treat the West so kindly as the last.
University 01 Detroit Mercy
NOTES
I For the full argument see my Technology's Story tellers: Reweaving the Human Fabric
(Cambridge : The MIT Press , 1985), Chapter 3.

2 Edwin Layton's ' Mirror Image Twins: The Communit ies of Scienee and Teehnology in
19th Century America" , Technology and Culture 12(4) (October, 1971), pp . 562-580 still
stands as one of the most insightful articulations of the differenee between scientific and
teehnological eommunities . Thus: "In the physical scienees the highest prestige went to the
most abstract and general - that is to the mathematical theorists from Newton to Einstein .
Instrumentation and applieations generally ranked lowest. In the teehnological eommunit y
the sueeessful designer or builder ranked highest, and the 'rnere' theorist the lowest" , p. 577.
See also Waller G. Vineenti, What Engineers Know and How They Know It: Ana/ytica/
Studiesfrom Aeronautica/ History (Baitimore: The Johns Hopkin s Press, 1990).
J David S. Landes, The Unbound Prometheus (New York: Cambridge University Press,
1969). Landes eharaeatures all eognition that does not take a rneans-to-ends form as
"superstition and magic", p. 21. See the following pages for further examples of Landes '
disjunetion between rat ionalit y and all other modes of eonsciousness whieh are defined as
defeetive by Landes ' dismissive, and oeeasionally derisive, tone .
The most artieulate eritique of this definition of rationality that I have yet eneountered is by
226 lOHN M. STAUDEN MAlER, S.l.
Renato Rosaldo , Cu/ture and Truth: The Remaking 01 Socia/ Ana/ysis (Boston: Beacon
Press, 1989). Thus, for example, "The point is to break object ivism's monopoly on truth
claims, not to throw out the bab y with the bath water. ... When the workings of culture are
reduced to those of a control mechanism, such phenomena as passions, spontaneous fun,
and improvised activities tend to drop out of sight."(p. 102) or "In my view, optionality,
variability, and unpredictability produce positive qualities of social being rather than
negative zones of analytically ernpty randornness ." (p. 112)but see also the full extent of his
argument throughout Part One.
• See, for some recent discussions of contextualism, Robert C. Post and Steven H. Cutcliffe
(eds.), In Context: History and the History 01 Technology, Essays in Honor 01 Melvin
Kranzberg (Bethlehem, PA : Lehigh University Press , 1989), pp , 150-171.
s Scholars representing a broad range of interpretative perspectives have come to agree on
the importance of contextual factors for interpreting science. See, for example, Arnold
Thackray, 'History of Science', in Paul Durbin (ed.), A Guide to the Cu/ture 01 Science,
Technology, Medicine (New York: Free Press, 1979); Evelyn Fox Keller, Reflections on
Gender and Science (New Haven , CT: Yale University Press, 1985); Bruno Latour, Science
in Action: How to Follow Scientists and Engineers through Society (Milton Keynes, Bucks.:
Open University Press, Cambridge MA: Harvard University Press, 1987); and Steven
Toulmin, Cosmopolis (New York: MacMillan, 1990).
• The concept of pro fessionalism , that ernpowers the licensed practitioner while
delegitimizing outsiders is a central achievement in 19th century Western society, See, for
example, Burton l. Bledstein, The Cu/ture 01 Professionalism: The Midd/e C/ass and the
Development 01 Higher Education in America, esp. Ch. 3.
7 My thinking here origina tes with the stud y of technological style in the United States. Thus
my application of these insights to European technological practice and to scientific practice

generally represents something of areach.
• Jererny Bentham, 'Panopticon; or, The Inspection Hou se', in The Works 01 Jeremy
Bentham (published under the superintendence of his executor, lohn Bowring), Vol. 4 (New
York: Russell and RusselI, 1962 (reprint of the 1838-1843 editionj), p. 39.
• I am indebted, for this turn of phrase , to Shoshana Zubo ff, In the Age 01 the Smart
Machine: The Future 01 Work and Power (New York: Basic Books, 1988), p. 320.
'0 Bentham articulated four defining principles:
(I) Cells " divided from one another ... secluded from all communication with each other,
by partitions";
(2) " Each cell has in the outward circumference a window, large enough, not only to light
the cell, but , through the cell, to afford light enough to the correspondent part of the
[inspector 'sl lodge. The inner circumference of the cell is formed by an iron grating, so
light as not to screen any part of the cell from the inspector's view" ;
(3) "To prevent thorough light, whereby ... the prisoners would see from the cells whether
or no any person was in the lodge, that apartment is divided into quarters, by
partitions" ;
(4) "Small lamps, in the outside of each window of the lodge, backed by a reflector, to
throw the light into the corresponding cells, would extend to the night the security of the
day." 'Panopticon' p. 40
It Works of leremy Bentham, Vol. 10, p. 226. Cited in Carolyn C. Cooper, 'The
Portsmouth System of Manufacture', Technology and Culture 25(2) (April, 1984), p. 193.
" Bentham, 'Panoptieon', p. 40 (author's emphasis).

13 Discipline and Punish: The Birth 0/ the Prison, trs. Alan Sheridan (New York : Random
House, Vintage Books, 1979), pp. 201-202.

14 Jean Hampton diseusses reeonciliation joined to eapital punishment. See her 'The
Retributive Idea ', in Jeffrie G. Murphy and Jean Harnpton, Forgiveness and Mercy (New
York : Cambridge University Press, 1988), p. 158.
15 Ibid., p. 19. Emphases mine
" 'Therapy, Not Punishment' , Harpers Magazine (August 1959), pp . 63-64 (emphases
mine). Menninger is quoted, together with Bertrand RusselI, B. F. Skinner, and Benjamin
Karpman to mueh the same effeet , in Herbert Morris, ' Persons and Punishment', The
Monist 52 (October, 1968), pp. 480-481. Morris ' essay has aequired the status of a classic
eritique of the therapeutic replacement of punishment. He roots his argument in the inherent
dehumanization of the person when guilt is replaced by treatment. " In th is [therapeutic]
world we are now to imagin e when an individual harms another his conduct is to be regarded
as a symptom of some pathological condition in the way a running nose is a symptom of a
cold." p. 480.
•, Harley Shaiken, Automation and Work in the Computer Age (New York : Holt, Reinhart
and Winston, 1983). Shoshanah Zuboff, In the Age 0/ the Smart Machine, passim .
•• See George F. Madaus, 'Curriculum Evaluation and Assessment', in P. Jackson (ed.),
Handbook 0/ Research on Curriculum (New York : Macmillan, forthcoming) . See also
Zuboff (especially Chs . 6-7 , 9-10) for discussion of the use of computerized algorithms in
decision-rnaking roles as precisely the same process of distancing the evaluated from the
point at which the evaluative judgment is rendered.
•• Erving Goffman, Asylums: Essays on the Social Situation 0/ Mental Patients and Other
Inmates (Garden City, NY, 1961). John F. Kasson, Rudeness and Civility: Manners in
Nineteenth Century America (New York: Hili and Wang, 1990). See Ch . I for his estimate
that an average of three new etiquette books annually before the Civil War rose to an average
of five or six per year from 1870 through World War I. See Chs . 4 and 5 on body control
advice .
20 See Lawrenee W. Levine, Highbrow Lowbrow: The Emergence 0/ Cultural Hierarchy in
America (Cambridge: Harvard University Press, 1988), p. 91.

2 . See Levine, Highbrow-Lowbrow, passim and Kasson, Rudeness and Civility , Ch . 7.
Levine notes that the gradual disciplining of audience interaction was accompanied by the
creation of separate establishments for upper dass and lower dass patrons.
22 On the early history of the wire news services see Daniel J . Czitrorn, Media and the
American Mind: From Morse to McLuhan (Chapel Hili , 1982), especially Ch . I; Richard
Sehwarzlose, 'Harbor News Association: The Formal Origins of the AP ', Journalism
Quarterly 45 (Summer, 1968), pp. 253-260; Robert Luther Thompson, Wiring a Continent;
The History 0/ the Telegraph Industry in the United States, 1832-1866 (Princeton:
Princeton University Press, 1947) and Richard B. DuBoff, 'Business Demand and the
Development of the Telegraph in the United States, 1844- I860', Business History Review 54
(Winter, 1980), pp . 459-479 and 1983: 'The Telegraph and the Structure of Markets in the
United States, 1845-1890', Research in Economic History 8, pp . 253-277.
23 The Molly trial was probably the first "national media" event in U.S. labor history.
Joseph Rayback describes the immediate and long-terrn effeets of the tr ial as folIows . "The
evidenee against them, supplied by James McParlan, a Pinkerton [detective] , and
228 JOHN M . STAUDENMAlER, S .1.
corroborated by men who were granted immunity for their own crimes, was tortuous and
contradictory, but the net effect was damning. All twenty-four were convicted; ten were
executed. The trial temporarily destroyed the last vestiges of labor unionism in the anthracite
area. More important, it gave the public the impression that miners in general were inclined
to riot, sabotage, arson, pillage, assault, robbery, and murder; and that miners were by
nature criminal in character and were to be condemned and disciplined by the more
respectable element in society. The impression became the foundation for the antilabor
attitude held by a large portion of the nation to the present day." A History of American
Labor (New York: MacMillan, Free Press, 1966) p. 133.
24 In his recent critical assessment of Richard Rorty, Christopher Norris uses U.S . media
coverage of the Irac war to exemplify some of the philosophical and pragmatic problems
that reside in electronically mediated public discourse. ('The "End of Ideology" Revisited :
The Gulf War, Postmodernism and Realpolitic', Philosophy and Social Criticism 17(1)
(November), pp . 1-40.
" Roland Marchand, Advertising the American Dream: Making Wayfor Modernity, 1920-
1940 (Berkeley: Un iversity of California Press, 1985), pp. 68-69 and passim . See also Daniel
Pope, The Mak ing oJ Modern Advertising (New York : Basic Books, 1983); Leiss, Kline and
Jhally, Social Communication in Advertising, and T .J . Jackson Lears, ' From Salvation to
Self-Realization: Advertising and the Therapeutic Roots of the Consumer Culture, 1880-
1930', in Richard Wightman Fox and T .J . Jackson Lears (eds .), The Culture of
Consumption: Critical Essays in American History, 1880-1980 (New York: Pantheon
Books, 1983), pp. 1-38. I am also deeply indebted to many conversations with Pamela
Walker Lurito for my understanding of changing advertising trends .
2. Debates ab out the effectiveness of advertisements in programming consumer motivation
are commonplace in recent studies. Michael Schudson, Advertising, the Uneasy Persuasion:
Its Dubious Impact on American Society (New York: Basic Books, 1984) argues the case
against it. Leiss, Kline , and Jhally Social Communication in Advertising, Chs, 2, 3, discuss
the pros and cons of both sides , citing Schudson, Stuart Ewen, Christopher Laseh, and
others; their own position tends to favor Schudson's .
21 (Robert O . Derrick, architect hired 10 design the Henry Ford Museum, Oral
Reminiscences) p . 50 [emphases mine) . Geoffrey C. Upward, A Home for Our Heritage:
The Building and Growth of Greenfield Viilage and Henry Ford Museum, 1929-1979
(Dearborn: The Henry Ford Museum Press, 1979), p. 50.
2. Otto Moog , German Engineer, in Thomas P . Hughes, American Genesis, p. 291,
author's translation of Otto Moog, Drüben steht Amerika: Gedanken nach einer
Ingenieurreise durch die Vereinigten Staten (Braunschweig: G. Westermann, 1927), p . 72.
Hughes cites another German engineer, Franz Westermann, saying "the most powerful and
memorable experience of my Iife came from the visit to the Ford plants .. . " (p . 292).
2. See Stephen Meyer , The Five Dollar Day (Albany: State University of New York Press,
1981), especially Chs . 5-8.
JO On the stockholder buyout see Alan Nevins and Frank Ernest Hili, Ford: Expansion and
Challenge: 1915-1933 (New York: Charles Scribner's Sons , 1957), pp . 105-111.

Nevins interprets the three resignations as folIows: "Ford ... looked aback with distaste on
the period of Couzen's activity in company affairs , when he had been unable to move freely .
The Dodge suit had of course intensified his desire for absolute authority. He was therefore
irritated by the presence of anyone in the company who might not work with hirn in complete
harrnony ." (ibid., p, 145).
31 See Geoffrey C. Upward (museum editor) , A Home for Our Heritage: The Building and
Growth of Greenfield Village and Henry Ford Museum , /929-/979 (Dearborn, MI: The
Henry Ford Museum Press , 1979) p. 76.
In 1919, Ford sued the Chicago Tribune for libel, was grilled on the stand with lines of
Questions demonstrating his flimsy educational background, and was awarded six cents in
damages. Robert Lacey, Ford: The Men and the Machine (Boston : Little, Brown and Co.,
1986), pp. 197-202.
J2 Greenfield Village was officially opened on October 21, 1929 when the aging Edison,
flanked by Ford and President Hoover re-enacted the invention by turning on a replica light
bulb . The " Festival of Light" was transmitted over a national radio network, one of the first
"live " media events in history. The only published account that treats the event in any detail
is Upward, Home for Our Heritage, Ch. 3.
33 Lacey, for example: " .. . there was only one beautiful room in the entire building: the
powerhouse. This was a spare , clean chamber which Henry had designed hirnself . .. . and he
created a very Ritz of power stat ions, all marble and gleaming brass dials and pipes. Around
the floor were set out little generators, raised on plinths like so many modern sculptur es ..."
pp. 149-50. See also, Collier and Horowitz, p. 71. See also Nevins, pp . 20-21.
The clean, uncluttered, "Ford" style that Charles Sheeler would make famous with his late
twenties photographs and paintings may represent the continuation of, and not a completely
fresh artistic reflection on, the Ford style. See Mary Jane Jacob, 'The Rouge in 1927:
Photographs and Paintings by Charle s Sheeler' , in The Rouge: The Image ofIndustry in the
Art of Charles Sheeler and Diego Rivera (funded by the Ford Motor Company Fund and
Founders ' Society Detroit Institute of Art s) and , more recently, Karen Lucic, Charles
Sheeler and the Cult of the Machine (Carnbridge, Harvard University Press , 1991).
3 4 Chicago Century of Progre ss International Exposition , Official Book of the Fair,
(Chicago: A Century of Progress , Inc., 1932), p. 11. I am indebted to Lowell Tozer , 'A
Century of Progress , 1833-1933: Technology's Triumph Over Man' , American Quarterly
4(l)(Spring, 1952), pp . 78-81 , for first calling my attention to the Exposition and to Cynthia
Read-Miller, curator of photographs and prints in the archives of the Henry Ford Museum
and Greenfield Village, for copies of the Official Book and photos of the iconography
referred to here.
For the Lohr quote see, Fair Management: The Story of a Century of Progress Expos ition
(Chicago: The Cuneo Press , Inc., 1952), p. 96.
3. Popular feelings about technocratic elitism were c1early mixed. Industrial unions
flourished in the Thirties as workers organ ized to contest managerial high-handedness . On
the other hand, even so shocking an episode as 1937's battle of the overpass evoked an
outpouring of fervent support for Ford's dictatoriallabor style in hundreds of handwritten
letters from ordinary citizens around the country . (Archives and Library Department,
Henry Ford Museum & Greenfield Village, Dearborn MI, Ace, 292 Box 43.) My cursory
inspection suggests that the vast majority of the letters, although not all, strongly favor
Ford's position .
3. Recently several sociologists of technology have developed the concept of "negotiation
spaces" to address much the same point. Thus : " ... we have shown how the proponents of
the project mobilized the actors in agiobai network and sought to create a relatively
autonomous negotiation space where a local sociotechnical network might be designed and
brought into being without constant interference from outside." lohn Law and Michael
Callon, 'Engineering and Sociology in a Military Aircraft Project: A Network Analysis of
Technological Change', Social Problems 35(3) (June , 1988), p. 290.
37 Steven Toulmin argues that it is precisely this systematic rejection of the cognitive validity
of the local, the oral, the specific, and the timely in favor of the universal, the written, and
the timeless, that characterizes what came to be the central orthodoxy of "modernity". For
his argument that aversion to the specific and local sterns from Europe's loss of nerve (and
sense of humour) in the face of the Thirty Year's War with its carnage of competing religious
orthodoxies, see Cosmopolis, Ch . land 2. The passage about the local, oral, specific, etc,
begins on p. 32.
David Harvey's interpretation of changing capitalist social definitions of space and time and
their influence on contemporary society is the most helpful and sophisticated that I have
read . See his The Condition 0/ Postmodernity (Cambridge, MA: Basil Blackwell, 1989),
especially Part III.
3. Shoshanah Zuboff records repeated instances where managers react "irrationally" to the
democratizing influences of open-access computer data bases despite the evidence that such
open access, a process for which she coined the name "inforrnating", increased efficiency
and profitability, managers experienced disorientation and fear when their control over
subordinates was threatened. See, In the Age 0/ the Smart Machine, passim
3. Peter Sand man has developed a persuasive model explaining the increasing social cost
that comes due when non-elites are excluded from such prioritizing debates. See his 'Hazard
versus Outrage in the Public Perception of Risk', in Vincent T. Covello, David B.
McCallum, and Maria T . Pavlova (eds.), Effective Risk Communication (New York:
Plenum, 1989), pp. 45-49, on the importance of inclusion of non-elites within such
prioritizing debates. See also Parker Palmer, The Company 0/ Strangers: Christians and the
Renewal 0/ America's Public Life (New York : Crossroads, 1985), for a cornplementary
societal analysis .
ARIE RIP
SCIENCE AND TECHNOLOGY AS

DANCING PARTNERS
I .INTRODUCTION
Derek de Solla Price, in an article in 1965, put forward the image of

science and technology as relatively independent, but closely interacting
activities; dancing partners as it were. I His analysis of the relation
between science and technology is still valuable, but he tended to look at
science and technology as separate, unified wholes, rather than ongoing
pro cesses and their interactions which cluster in various ways and are
labeled "science"and "technology", also in a variety of ways. If this point
is added to his analysis, it allows us to raise, and to some extent answer,
further important questions: about the patterns of the dance, the contexts
in which it occurs, and secular changes and transformations.
Historians, philosophers and sociologists will immediately agree that
there is a complex relation between science and technolog y, which should
not be reduced to simplistic formulae like "technology is applied science" .
What one should say instead is less clear. Price's image still suffers from
a certain reification of "science"and "technology" , and is thus
insufficient to counteract the danger that idealtypes of "science"and
" technology" are posited, based on a few examples, or just on the
projection of the ideas of the particular philosopher, after which the
analysis is conducted in terms of these idealtypes. The evolving character
of science, of technology, and of their relations should therefore be
foregrounded. Historians will immediately agree, but they may run the
opposite danger: getting lost in particular historical episodes and
particular individuals and social interactions.
To navigate between the Scylla of philosophical idealisation and the
Charybdis of historical particularism requires more than a recognition of
the two opposite dangers and the resolution to avoid them . One needs a
perspective that overcomes the limitations, at least in principle. The en-
trance point into the complexity ofthe real world of scienceand technology
taken here is to view both as search processes which gradually become
embedded in local practices as weIl as in more cosmopolitan fields of
science or technology , and which have at least some non-local outcomes.
231
P. Kroes and M. Bakker (eds.), Technological Development and Science in the Industrial
Age, 231-270.
232 ARIE RIP
The non-local or cosmopolitan aspect of science and technology is not

automatieally given. Science is not universal by definition, it takes a lot of
effort to decontextualize local findings,? And technology is not
automatieally applicable in other situations, it requires work, both on the
technology and on the situations, to create a functioning technology. 3 The
social side of this process of decontextualization, whieh is also a
recontextualization into "fields" of science and of technology, consists in
exchange between practiees, circulation and network creation. The
mobility of the Renaissance scholars and engineers," of 19th century
colourists," and of 20th century physicists and biomedieal scientists," were
essential conditions to the creation and maintenance of a field, and of
scientific and technologieal identities of the practitioners. 7 The cognitive-
technieal side of the process is the emergence of shared disciplinary and
professional repertoires. For science, this process is clearly visible since
the late 18th and early 19th century. For technology it starts by the middle
of the 19th century (but with important precursors, as in building of the
cathedrals and fortifications in the Middle Ages) ."
A number of implications can immediately be drawn from this
perspective. One is the possibility of a certain mutual adaptation and
division of labour between science and technology: thus, dancing
partners. The dance, however, is not pre-given, but is created in particular
historieal circumstances and made possible by particular "cognitive
infrastructures". For example, the emergence of "technieal models", like
the model of an electromotor (cf. also the late 19th-century work on a
general machine), and unit operations in chemieal technology, allows
scientists and engineers to do their own thing and still relate to each
other.?
A second implication is the importance of anticipations on the value of
expected outcomes as on the one hand incentives for action, and on the
other hand as determinants of what survives in science and technology.
The dancing partners must anticipate the next steps, and find them
important, to be able to continue to dance . More concretely, economic
and strategie considerations come in not only after science and technology
have produced their products (knowledge claims, blueprints and
prototypes), but are part and parcel of the production process.
Enlightened econornists'? have, by now, developed such "endogenous"
analysis of technology (and may perhaps extend their analysis to science),
but sociologieal and philosophieal extension of their work is essential for
a full understanding of the dance.
SCIENCE AND TECHNOLOGY AS DANCING PARTNERS 233
Thirdly, it must be possible to trace changes in the relations between

science and technology, transformations even, in these terms in the past
and in the present. One may, in fact, speculate ab out a new kind of
intertwining of science and technology that occurs as part of a conscious,
strategie mobilization of cognitive -technical potential. This adds up to a
strategie or politieal transformation of the science-technology complex,
including the emergence of other organizational and institutional forms
than we are used to now (for example, a very different type of university).
The dance halls themselves change .
In general, search processes and outcomes in science and technology
are shaped by the dance halls, the musie, and the dancing patterns that
exist. The dancing itself creates new patterns, however, new music is
written, and the dance halls may weIl be modified in the process .
2 . THE ROLE OF SCIENCE: AGAINST THE LINEAR MODEL
It is convenient to start with abrief discussion of the so-called "linear

model" of the relation between science and technology. That technologieal
innovation derives from scientific discovery, as it were in a linear sequence,
is a myth, but aprevalent myth. As a myth it is tenacious because of its links
to important legitimations of science as the horn of plenty, and of
technology as the magie wand . And one indeed can find examples where it
makes sense to relate technical innovation to a preceding scientific
discovery. The linear model has some truth in it, but it hides more than that
it helps our understanding. It is important to replace it by an alternative
model, even if one might argue that one should not try to capture the
variety of science and technology in a single scheme, to avoid creating what
is really just another myth. Models are still important, as long as their link
with ongoing processes is made clear , rather than that they primarily
depend on myths about the nature of science.
The myth of science as the source of all sorts of good things is brought
up in public pronouncements by scientists and others, and is especially
striking in the image of the goose that will produce golden eggs as long as
it is fed properly. One can trace back this notion for more than a hundred
years, and find quotes like this one from Helmholtz in 1862:
The scientists - for the benefit of the entire nat ion and almost always at its request and
expense - are seeking 10 multiply the knowledge which can serve the increase of industry,
234 ARIE RIP
wealth, and the beauty of life, the improvement of the political organization and the moral
development of the individual. Yet, not immed iate utility must be looked for, as is so often
done by the uninformed. Everything that informs us ab out the natural forces or the forces
of the human spirit is valuable and in time may pro ve useful , normally in a place where one
had least expected this .!'
To continue with the metaphor: the geese should be allowed to range

freely in the meadow of science, and one might have to seek for the eggs
in the dung heap. The point is not that the geese of science never lay
golden eggs, only that reliance on such an umbrella legitimation hinders
our understanding what actually happens .
The same point can be made about another myth prevalent in our
society: technology as the means to achieve, supposedly in an
unproblematic way, whatever we want. What Mankind Can Dream ... ,
Technology Can Achieve - this is the slogan of Fujitsu Company: even
Japanese companies are using this image to spread the message of
benevolent high-tech. The reason why technology should achieve so much
is often an implicit linear model: the stereotypical image of technology as
the application of all-powerful science, combined with the mistaken
equation of the very real and increasing role of science in modern
technology with causal sequence from science to technology. 12
Technological innovation is not necessarily the resuit of a trajectory
starting with scientific discovery - take as an example the zipper, a unique
invention, for which neither the original 1891 patent, nor the first
practicable design of 1913 owed anything to scientific research .'" When
the innovation is related to science, there is always a lot of work to do
(including intervention of and response to external circumstances) before
a scientific discovery results in a working process or product. Again,
myths about the power of science abound, for example in the way
Alexander Fleming is honored as the discoverer of penicillin, while it is
through Howard Florey's and Ernest Chain's work (and that of many
others) that we have penicillin as a product." In fact, what Fleming
discovered could best be described as a mould extract that kilIed certain
kinds of bacteria.
As a more extended example, the case of linear polyethylene can be
used to show something of the real-life pressures and contingencies."
After the discovery by Ziegler (and others) of catalysts that produced a
regular polymer at low temperatures and pressures, there was stilI very
much work to do. Patents were taken out, and Iicenseswere obtained, but
these were very much "hunting licenses": " .. . everyone went through
much the same traumatic adolescence in developing product application
information." The main problem was that the new polymers fulfilled
their promise too weIl:
As soon as enough linear polyethylene became available to try to put it to work, it was found
that the "Chain Straighteners" had done their work too weil, for in polyethylene as in
society, absolute straightness turns out to be more virtue than is wanted for practical
purposes. The advantages of hardness, stiffness, strength and heat resistance in finished
products were gained at the cost of seriou s difficulties in the manufacture and use of those
products. Since Phillips polyethylene had the "purest", 100070 linear cha in structure, it
suffered these difficulties in the highest degree . {... )'6
In every one of the major developed markets for which linear poly-
ethylene seemed suited, the initial bright promise was tarnished by one or
other unanticipated "different" characteristic that proved to be a
significant disadvantage. For example, the established markets for
packaging film were not open to linear polyethylene because its high
crystallinity rendered it opaque and prone to splitting .
All these nightmarish problems came to light just as the first polyethylene plants were
coming on-stream, with all the trauma associated with plant start-ups, One after another,
each of the "rniracle progeny of plastics " began to look more Iike a retarded child . The
company managements that had fathered them were forced to contemplate infanticide and
were, in turn , threatened by their boards of directors with financial sterilisation. Heads were
greying, if not rolling, in more than one organisation, and PhilIips, for one, were very close
to closing their plant and aborting the whole enterprise when salvation showed up in the
form of a toy .
It would be gratifying to a chemist to be able to report that the research teams had leaped
to the rescue just in the nick of time with technical solut ions to all the technical problems.
As a matter of fact, they did eventually solve nearly all of them , but not quickly enough to
save the situation had it not been for the fortuitous burgeoning of an unprecedented,
unplanned, unforeseen market. The Wham-O Toy Company introduced the " Hula Hoop" ,
made from extruded polyethylene tubing , and the ensuing craze created a non-critical
demand that swept the glutted warehouses clean and put the plant s back on an around the
clock production schedule.
Th is fad, even though it faded rapidly, bought time for the polymer chemists to
determine that most of their woes arose from the super-crystallinity of their super-linear
polymer chains, and that by judicious adulteration with touches of a second monomer
(introduced in the polyrnerisation step) , they could put a very occasional kink in the
molecular chains and thus alloy the crystallinity of the polymer just enough to suppress most
of its deficiencies while preserv ing most of its sterling qualities. {... )17
236 ARIE RIP
In this brief example, many features of ongoing processes of science and

technology are already visible. The case is also an illustration of a two-
branched model that should replace the linear model as a more realistic
view. Critical reading of historical case studies and a good look at ongoing
interactions suggests a two-branched model, with an empirical or semi-
empirical finding as the starting point, to be appropriate. Taking the idea
of a "finding" as an unanalyzed category for the moment," and as the
source of the developments, two different kinds of activity are seen to
branch out from it:
1. Exploitation (technological development, pilot process, feedbacks),
and
2. Exploration to increase understanding (through scientific research).
The insights derived from the exploration branch may sometimes be
called in to assist and improve exploitation (trouble shooting, rationaliz-
ation, and what can be called "transformation of the exemplar", see the
example of synthetic dye chemistry in the next section). The relationships
are depicted schematically in Fig. 1.
~~ ~
V
• EXPLOITATION
(Semio)emPiricalQ
~~
- .....
findin~ '::- ': \. rauonalization
rec ogmzed as such ' \ "- .;:.>
~--
~~
\
~
.
~CIENTIFIC EXPLORATION
Fig. I. The two branches of development of a finding .
The history of the linear polyethylene case further illustrates this model; 19
another example is the rapid exploitation of Röntgen radiation after its
discovery in the laboratory. While the original starting point may weIl
derive from scientific work, as these two examples indicate, it clearly is
not a matter of applying a scientific finding. There appear to be three
types of connections between science and technology: a laboratory effect
or method is exploited for another purpose, as happened for example with

recombinant DNA techniques or with hybridomas. Or a new domain of
nature is opened up in the laboratory, and then also available for technical
exploitation; think of Röntgen radiation, of nuclear fission. Thirdly, as I
will discuss below, science may be a source of powerful heuristies for
technologieal search processes.
When the linear link between science and technology is severed, one
becomes sensitive to the limitations of further received views. For
example, knowledge transfer should never be conceived as sending and
receiving. There is "local innovation" rather than adopticn.i" And even
without such active adoption, there will still be mutually complementary
processes of aggregation of user needs, and disaggregation of
"universal", i.e. laboratory-based, science findings."
The analysis of the role of science in this section is not exhausted by
what was said about the two-branched model. There , the focus is on a
specific finding and later developments, rather than on the ongoing
activities of doing science and doing technology. What is missing
specifically is the fact that scientists and technologists, but also other
actors like industry and government, expect that there are productive
relations between science and technology, and attempt to realize these, for
example in innovation-oriented R & D stimulation programs . While these
efforts to implement a strategie science and technology policy are a
striking feature of the science and technology landscape of the 1980s and
1990s, they are possible only because of much longer-term processes of
co-evolution of science and technology.
Analytieally, this implies that one should look at science and
technology at three different levels: one level is that of legitimations and
reified notions of science and technology, which were criticized for
obscuring the relation between science and technology in developing new
findings. Going down to the basie level of practiees of science and
technology, one can see patterns like the two-branched model. This is a
specific pattern, in relation to new findings and their exploitation and
exploration. There are also patterns, like that of decontextualisation and
recontextualisation to fields of science and technology at a third, and in a
sense intermediate level. It is at the intermediate level that recent
transformations, signalIed by the new science and technology policy
activities, occur.
238 ARIE RIP
3. SCIENCE AND TECHNOLOGY AS SEARCH PROCESSES
The complexity of the world of science and technology can be addressed

from the perspective that the practices of science as weil as technology
consist of (contextualized) search processes. That is, there is perhaps no
essential difference even while their historical forms are separate." As
institutions, science and technology are different. The products will often
differ (working experiments and texts versus designs and blueprints,
artefacts, pilots), and the contexts are still different, even though they
used to be more different than they are now. 23
One implication of the search perspective is that the concrete links
between science and technology will be extremely varied, as they derive
from the brico/age of the (rejsearchers." Historically, one can see both
separation and intertwining, e.g. the temporary interaction in the
Renaissance (and in a more imperialistic way, in the 17th century), the
new interactions and demarcations in the 19th century (including the
invention of the label "applied science" !), and the high tech and high
science world of the late 20th century.
3.1. Search Heuristics and Quasi-Evolutionary Development
As the economists Nelson and Winter have emphasized, search processes

are structured by heuristics, that is, rules that do not guarantee, but
promise some, success in problem solving," Such heuristics are not
individual (as studied by psychologists interested in problem solving), but
shared heuristics, embedded in organizational contexts and/or
practitioners' communities. The products of the search processes are
exposed to assessment and selective use by others . So there is variation
and selection, the basic building blocks of evolution.
But it is quasi-evolutionary development, because intentional and
strategic: heuristics anticipate on selection, and selection environments
are modified by search actors. Often, expectations about success, or
promise, of certain routes or options, in relation to perceptions of the
environment, are the basis for decisions and directions taken." Such
expectations may become stabilized (in an organization, in a professional
community, in a sector of society or society as a whole - think of the
recent promise of biotechnology), and form a cu/tura/ matrix 0/
expectations." It is always possible that expectations remain informal,
and barely articulated, but with the increasing need to justify and
legitimate actions, e.g. of an R & D department with respect to its board
of directors, or technologists asking for government support, there is ad
hoc or even systematic articulation (including special assessments of
trends and potential of scientific and technological domains, produced by
consultants or expert committees).
So there are two processes at work: one is the articulation of
expectations and incorporation of new information and assessments; the
other is sedimentation, the emergence of a shared and stabilized repertoire
of expectations , on which actors can draw as aresource. As soon as that
happens, one can speak of a matrix of expectations, because decisions,
choices and further assessments will be embedded in it. Such matrices of
expectations are never static; expectations evolve, and there is negotiation
about their worth (their reliability and appropriateness). So any
description of a matrix of expectations is a snapshot, taken of an evolving
process. Still, it is important to intro duce the concept as standing for a
relatively stable phenomenon, because it allows one to understand, first,
how heuristics can spread and how commitments to them emerge, and,
second, how others than search actors can be involved: even if not actually
participating in problem solving, managers and third parties will base
decisions drawing on the cultural matrix of expectations.
Because such matrices of expectations play an intermediary role, it
becomes understandable in principle how the whole development is
structured through rules and institutions formed around expectations,
even before there is any actual external selection .
I shall use the example of synthetic dye chemistry in the 19th century
to briefly illustrate the dynamics, and prepare the ground for a discussion,
in the next sub-section, of socialloci of search processes. 18 The success of
aniline red (discovered/invented in 1859, after a variety of attempts
comparable to Perkin's "extraction" of aniline purple from aniline)
attracted other producers into the synthetic dye field, and the correspond-
ing, and rat her primitive heuristic was "tinker with aniline", a heuristic
that indeed produced a number of interesting dyestuffs. The aniline red
paradigm was rationalized by Hofmann on the basis of the chemical
theory of types, and although his type formula did not represent a
definitive solution as to the chemical constitution of the "mother
compound" rosaniline, it was quite effective in stimulating and directing
innovative activity. 19 The attempts at elucidation of the chemical
constitution, and the perceived success of such attempts in guiding search
240 ARIE RIP
for further dyestuffs, can be seen as the first step toward a new, more
general heuristic : "Look for the Muttersubstanz (that is, the skeleton of
the dyestuff's molecule), and if it is found, each derivative of the
Muttersubstanz may be an innovation" . This heuristic became the central
component of a new paradigm of purposeful synthesis, which could
already be discerned in rudimentary form in the elucidation of the
chemical constitution of alizarin and its subsequent synthesis in 1869.3 0
C1early the two-branched model of development of a finding is
applicab1e, but in addition, search heuristics of more general va1ue
emerge. In these examples, the heuristics cluster around an exemplar: a
product or process that is an exemplary achievement. Thomas Kuhn
(1970) redefined his "paradigm" as consisting of such an exemplar in
combination with a " disciplinary matrix" . In the same vein, we can now
speak of a technological paradigm as an exemplar (in this case, one
achieving technical and commercial effects), embedded in a cultural
matrix of expectations. Further work in the frame of such a paradigm
leads to a trajectory, a sequence of innovations that are related through
their use of a cluster of heuristics (which may weIl be modified along the
trajectory);" Note that the analogy with science remains : "normal
science" in the framework of a paradigm, or, alternatively, the pro-
gressive elaboration of a Lakatosian research program, are equiva1ent to
the technological trajectory. In fact, the chemistry in that period can be
analyzed with the same categories of search processes, heurist ics and
partial stabilization in paradigms.
In another example from synthetic dye chemistry, the emergence of
azo-dyestuffs, the dynamics of emergence of a technological paradigm
can be seen in detail. There was an exemplary process (a coupling
reaction, named after Griess, which produced coloured compounds that
were interesting to scientists); there was an exemplary product (Roussin's
naphthalene dyes, with very useful properties, found by trial and error in
an industrial workshop), both with their own search histories and kept
more or less secret. The actual matrix of expectations emerged in a
dramatic way, with Hofmann disclosing the secret of preparation of both
classes of compounds in 1877, and emphasizing that a vast domain for
synthetic dye chemistry was now opened, of which the limits were not yet
in sight.
The conditions outlined here for the emergence of a paradigm are
important for ongoing technologica1 development in general, even when
these do not fit the model of a paradigm plus trajectory."
3.2. Social Loci
Whether sedimentation into a paradigm occurs or not, expectations create

an agenda for research directions and investments, mobilize actors and
draw them together. In a sense, an arena is created where actors have to
take particular expectations as weIl as actions by other actors into
account, because these are relevant for their own assessments and
decisions. An example is the emergence of a "technical community" of
practitioners, who are focussed on the same, or on mutually relevant
issues, and interact directly or indirectly,33 The 19th century colourists
were weIl known to each other, and relatively independent of employing
organizations." They circulated from one firm to another, and often
commanded high salaries . The colourists carried the technical heuristics
of dyeing and making appropriate dyestuffs, and were thus the social
locus of these heuristics.
In other cases, other actors and institutions were involved . While the
paradigm plus trajectory of rosaniline dyes, for example, were carried by
university researchers as weIl as industrial colourists (there was no
specialized locus yet), the azo-dye trajectory coincided with the emergence
of the industrial research laboratory and the routinization of research."
These examples can also be used to make a further point: whether it is
a technical community, institutionalized research laboratories, or some
further variant, a sociallocus will introduce a dynamic of its own, because
of its social and economic make-up, and its interest in survival. Further
search processes will be started, and expectations will be voieed, to
maintain the power of the sociallocus and legitimate it. This aspect of the
dynamic is captured rather weIl by the actor-network approach of Callon
and Latour, because the arrow of technological dynamics here goes from
network processes and strategie interactions to technologies.:" Particular
heuristics can become so entrenched in a social locus, that they
"determine" the course of developments, and the approach to new
problems.
In general, one can predict that there will be a phase of rather un-
directed search and heterogeneous expectations, a phase of sedimentation
of heuristics and expectation matrices and their institutional embedding;
and a phase of elaboration of more or less accepted heuristics and
expectations. This prediction can be related to the successful approach of
Abernathy, Clark and others in terms of a design hierarchy and its
maturation." Note that a design hierarchy has a sociallocus: it is carried
242 ARIE RIP
by the producer-consumer network called an oligopolistic market. In the

same vein, Dosi proposes a two-phase model of technical change and
industrial structure: a Schumpeterian trial and error phase, and a phase of
oligopolistic maturity. 38 As with the authors quoted by Dosi, it is not clear
if the model is descriptive, or has a normative point, e.g. should one
stimulate oligopoly when a technology and industry is "mature"? The
proposal is clearly important, but the paradigm is reified into an
independently existing factor, about which questions can be asked as "In
what phase of development is the paradigm of domain X?", the answers
to which would then have policy implications in a mechanical way. I
would argue that a process analysis of the technological dynamics is
necessary in which questions about paradigms can, and should, be
embedded."
A final point to be made is that institutions relevant for technology
(and similarly for science) emerge and function as a response to the
activities of anticipation on selection (and on strategies of other actors)
and as an attempt to reduce uncertainty and create stability in the
environment. Van den Belt and Rip introduced the notion of institutional
nexuses here, that is, particular linkages between variation processes and
selection environments, that can be adopted by others:" Examples of
institutional nexuses are the patent system and the test laboratory (where
findings are subjected to what one could call a simulated selection
environment). The spread of such a nexus is clear in the case of routinized
industrial research to exploit inventions, which emerged in the late 19th
century in electrotechnical and chemical industries, and was a necessary
carrier for successful heuristics of system design, and rational synthesis,
respectively.
Recently, and especially with the rise of strategic mobilization of
science and technology, a new type of institutionalization has emerged,
which allows virtual selection on the basis of expectations, as it were
before the fact. Expectations then become aresource, and appraisal and
negotiation routines develop. Biotechnology provides a clear example of
such new institutions. Biotechnology R & D firms operate by creating
expectations (up to rapid increases of their stock market value after
introduction of their shares on the stock market)", but have to keep up
these expectations by doing the right things (engineering a growth
hormone, or insulin, or interferon) and surrounding them with the right
publicity. These may not even be the most profitable innovations (which,
especially during the first years, were drugs and feed components in
animal husbandry), but they are necessary to maintain visibility for wider
publics. The whole business is precarious, as the decrease in stock market
value of the same firms shows, as weIl as the difficulties into which a
number of them have run .
More generaIly, one could say that in a world where innovation
competition becomes more important, early promises about innovations
will be valued as such, that is, as building blocks of the necessary matrix
of expectations . A speculative market of early promises emerges; a
process fuIly analogous to the emergence of stock exchanges, where
shares in firms (and before that, in merchant ventures) could be traded in
terms of the expected profits. The R & D firms that carry this market
started out as a specific nexus, but when this type of firm became
institutionalized, the market became relatively autonomous, and
academic research groups and other actors could also start offering their
early promises for "sale", i.e, to be taken up in exchange for resources.
Strategie mobilization of science and technology can then become a
regular phenomenon, rather than ad hoc and based on bilateral
negotiation.
Three points can be made to conclude this outline of a quasi-
evolutionary model of search processes in science and technology.
Scientific and technical communities used to be the locus for appraisal of
expectations and developments of heuristics as such, and while being
shaped by institutionallocations of the practitioners (as will be discussed
in the next section), could do so relatively autonomously. Many more
actors are now involved explicitly. Analytically, the point is that the
activities continue to be at the level of fields of science and technology
(including patterns of decontextualization and recontextualization, as I
noted in Section 2).42
A second point is that the model connects two aspects that are often
kept separate: the action and process level, where actual production of
technology, adoption and diffusion occur, and the level of cognitions and
legitimations, where heuristics, expectations, and rules are to be found . In
other words, to "producing a working artefact" it adds two other
important activities: "making a clever move" (this is getting more
attention in recent technology studies)? and "telling a good story". The
telling of stories, so that they have effects, is an important aspect of socio-
cognitive-dynamics, but has been neglected in general (even if there is
some interest from organization studies), but certainly in technology
studies."
244 ARIE RIP
Thirdly, for sociological and economic studies, the model outlined

here is more general than the traditional focus on firms (especially by the
economists) and on technical communities (especially by sociologists and
historians). It also adds to innovation studies , as weIl as Latourian studies
of Machiavellian engineers, which neglect how actions, interactions,
adoptions etc. add up to something at the collective level that is
sufficiently stable to be drawn upon by actors as aresource.
4. DIVISION OF LABOUR BETWEEN TECHNICAL PRACTICES

AND ACADEMIC RESEARCH ON THE BASIS OF
TECHNICAL MODELING REPERTOIRES"
Engineering design at the locallevel is a socio-cognitive process whereby

specific goals or mandates guide the manipulating of properties and
configurations of the artifacts-to-be-made, but such design processes also
relate to a more or less cosmopolitan design culture, which contains
technological paradigms, routines, heuristics, norms and standards. For
the question of relations between technology and science the notion of
"technical model" is important. It is a reasoned (though not necessarily
fully articulated) conceptual representation of dass of artifacts and
artifact systems that is present in design practices, in technical
communities, and (after a time) also in technical education.
Technical models, like the general model of a bicycle, an electromotor,
a bioreactor, or a membrane, are mental or material representations of
artifacts as a system of interrelated and mutually constraining sub-
elements. The importance of such a model is that it allows designers to
infer aggregate artifact behaviors from specific element parameters, the
overall configuration of elements and the mutual constraints obtaining
among them . It should be noted that any given species of artifacts may be
modelled in a number of different ways, depending on the specific
behaviors which are of interest. Thus, for example, ships' hulls may be
modelled for their hydrodynamic, their aesthetic, or their structural
properties; electrical power plants may be modelled as self-contained
thermodynamic systems or as self-regulating components in a larger
system of variable demand and power loads.
Thinking in terms of, and working with, technical models enables
designers to engage in virtual manipulation of design parameters with an
eye to optimization along relevant evaluative dimensions. In addition,
they provide a potential cognitive infrastructure for the participation of

actors-at-a-distance, in an organizational and strategic as weIl as
epistemological sense . Improvements in ship hull design can profit from
work on the hydrodynamics of particular propellers in one research
institute, and the development of new welding techniques in a specific
firm, because they all relate to each other in a technical model.
Within weIl developed technical domains , technical modelling
repertoires are routinely available to designers and will be incorporated
into engineering curricula, complete with standardized symbol systems
and application protocols. The subjective element here is reduced to
decision making about how to apply or adapt available modelling
strategies to local design problems, e.g. how to adapt standard practices
for modelling ships' hulls to the modelling of this particular ship's hull ."
HistoricaIly, such technical modelling repertoires emerge out of local
modelling, and its improvement. Consensus among local designers on
efficient nomenclatures and symbolizations and on the relative salience of
specific relationships and parameters in technical models for particular
types of artifacts, are important steps toward such a repertoire. It is then
institutionalized, and further developed within professional discourses
and engineering curricula.
When technical models are part of the shared repertoire, it becomes
possible to improve them as such, independent of particular design tasks .
This can be seen clearly in attempts towards a theory of the general
machine," and allows academic research to contribute to technological
dcvelopment (at least in principlc). 1 shall call such work on technical
models "meta-modeling" .
4.1. The Globalization 0/ Technical Modeling
A local technical model can be problematized because it fails to solve

design problems, for example in achieving more ambitious specifications,
and this may lead to local excursions into meta-modeling, sometimes
involving scientific research coupled with the importation of professional
researchers, as weil as to efforts to learn from the experiences of others,
including appropriation of published findings and other forms of
information exchange with technological competitors. Within particu-
larly large and wealthy state agencies or private companies such meta-
modeling may become a differentiated organizational activity, either on
246 ARIE RIP
an ad hoc basis to solve emergent design problems and appropriate new

technologies (e.g. in the form of incidental experimentation within such
organizations as Army artillery corps, railways, national public works
agencies) or more structurally as indicated by the emergence of state and
corporate research laboratories at the end of the 19th century and again in
the 1920s and 1930s. Meta-modeling then can be seen as fundamental
research, although not necessarily connected to basic science as
traditionally viewed.
It is true that technical models are nowadays often anchored in basic
science, in the sense that design-relevant representations of artifacts draw
heavily on scientific theories about phenomena which are embodied in the
artifacts, rather than starting out as local ad hoc constructions. But even
then, the construction and optimization of technical models, i.e. meta-
modelling, remains a distinct type of activity, and the province of research
specialists within a global di vision of technologicallabour.
The configurations of actors in this division of design labour are
nationally and historically contingent, but a dominant pattern emerged
during the nineteenth century, at least in the nations of the European
continent, which is still visible today. Professional engineering
associations and polytechnical schools gradually insinuated themselves
into loeal design proeesses by first establishing the neeessity and
legitimaey of abstract, theoretical, and therefore generalizable st yles of
technical modelling;" and subsequently assuming some of the burden of
the associated labour of meta-modelling. In later phases, universities,
particularly in fields Iike medicine and chemistry (e.g. organic chemistry
in relation to synthesis) played such a role.
There was a strong social dynamic to the emergence of the dominant
pattern. Unanimity, and uniformity of design protoeols (thus of technical
models) was a central value for the newly self-conscious engineering
professions. It legitimated their claim to a unique role in design practices
by their demonstrative commitment to critical, "scientific" scrutiny. In
the second place, uniformity of technical models, including the
standardization of symbolic representations and algorithms, facilitated
communication among professional engineers (e.g. electrical circuit
diagrams or stress calculations for bridges produced in one location could
be routinely read by professional colleagues elsewhere). As uniformity in
technical modelling thus cemented professional solidarities, it also
significantly set up barriers against competition from non-professional
practitioners.
Teaching staff at the new polytechnics had in common their mandate

to train experts capable of producing state of the art designs and therapies
in a variety of technical situations. This meant that they could not
inculcate rote solutions to invariant problems but were compelled to
impart more general representations which could guide optimal design
strategies in a variety of settings, i.e. to teach technical models rather than
techniques. This acted as a spur to faculty to produce (or at least explicate
and codify) the requisite technical models, including fundamental
research to establish salient relationships and parameters involved. In
addition, engineering professors traditionally feIt (and may still feel)
forced to academize technologies as part of an ongoing collective status
struggle with the classical universities. So they oriented their research not
only to the specific design needs of technologically committed
organizations but also to participation as scholars in emerging disci-
plinary fields (i.e. to profiling themselves as professors of engineering
science). Being institutionally isolated from the specific local design
imperatives of state agencies or private enterprises and seeing the
education of engineers as their major task, they were both enabled and
compelled to focus on the basic and universal aspects of the design
process, including codification, standardization, and the formulation of
general technical models."
Note, however, that the engineering faculty did not fully monopolize
the role of agent of cosmopolitanization in the division of design labour:
engineers associated with the various technical corps within the military
and the large state infrastructural agencies which were part and parcel of
nineteenth century state formation in Europe also continued the
originally French tradition of contributing to the cosmopolitan fund of
technical models .
Cosmopolitanization occurs in different technical domains at very
different points in time. Nonetheless, both the academic motive and the
institutional framework for design cosmopolitanization is given with 19th
century educat ional modernization and the rising status of "science" as a
legitimate academic pursuit. From that point on, the division of design
labour and the emergence of cosmopolitan design networks becomes a
more or less standard milepost in the development of new technical
domains . Synthetic organic chemistry passed through this phase - and in
a number of steps - in the last half of the nineteenth century (and a replay
occurs in the 1970s and 1980s with "synthon" theory) . In chemical
technology, the emergence of unit operations in the first decades of the
248 ARIE RIP
century and the effort to introduce chemical technology/chemical

engineering into the universities, is a clear example. Among civil
engineering technologies, academic involvement in the mathematical
modelling of reinforced concrete construction became, from the late
1890s on, a logical extension of traditional professorial involvements in
the theory of applied mechanics and elasticity in general. 50 Classical
biotechnology was a late bloomer, showing cosmopolitanization by the
1940s and 1950s and, from about 1960 onwards, attempts at academ-
ization, e.g. the founding of university departments, the publication of
handbooks and the founding of scholarly journals." There are other
examples of late (and sometimes negligible) entrance in academe, e.g.
polymer science and engineering, heterogeneous catalysis (pioneered by
the oil companies and some big chemical firms)," and atomic energy
engineering.
What is important for my overall argument is that there is a pattern,
and that it cannot be specified in terms of a relation between "science" as
such and "technology" as such. Rather , it is an autonomous development
of technical domains, which makes them resemble scientific domains . In
fact, the distinction becomes tenuous, and seems more related to the
nature of the links with local design work, than to any inherent
characteristic of science in contrast to technology .
4.2. Additional Linkages between Academic Research and

Technical Work
In the early twentieth century, the institutionalloci for the cosmopolitan

design networks, were the polytechnics and technical universities and the
big industrial and state-related research laboratories. With the notable
exception of the United States, the classical universities did not want to be
seen as centers of applied science research. 53 The polytechnics (acquiring
formal academic status, i.e, the ius promovendi, in Germany, Austria,
and Holland around the turn of the century) had effectively monopolized
engineering science in most fieIds. Thus, the electrical and electronic
revolutions, despite their historical roots in the physical sciences, did not
become an occasion for university-based specialties because of the
institutionalization of electrical engineering at the technical schools (and
of research in state and corporate R & D labs). Until after the turn of the
century, academic geology anxiously kept its distance from polytechnic-
SCIENCE AND T ECHNOLOGY AS DANCING PARTNERS 249
based mining engineering, concentrating on classifications and the

etiology of formations rather than on technical models directly relevant to
the discovery and extraction of ore deposits. A somewhat different
relationship prevailed between university chemistry and the chemical
technology which had emerged as a pendant to various process
industries .S4
The definitive breakthrough of industrial research laboratories
happened by the 1920s(state laboratories had a longer tradition, but were
originally related to the standard setting and regulatory tasks of the state).
It was increasingly often the case that ideas for new artifacts and designs
emerged out of local industrial meta-modeling research itself. For
example, Philips company's facile entry into radio vacuum tube research
after the first world war was a direct consequence of the knowledge gained
from meta-modeling work on medical X-ray tubes du ring the war.
Within this set-up, new possibilities for Iinkages between academic
research and technical work (local as weIl as cosmopolitan) were
emerging. The simplest way to show this is to start with the technical
hierarchy in design:
• components (e.g. materials , nuts and bolts , resistors and condensers,
radio tubes) that do not "work" by themselves, but have to be as-
sembled;
• devices (e.g. a pump, a switching circuit, a sensor) that are assembled
sufficiently to show their primary effect;
• artifacts (e.g. a machine, a bridge, a radio), that work by themselves;
• systems (a plant, an electricity network , radio broadcasting plus
receivers plus organizations to produce radio programs) that fulfill a
function.
The levelsmerge into each other (a radio tube can also be seen as a device),
but the point is that a hierarchy exists, and that design work can take place
at different levels.ss At first (analyticaIly, but often also historicaIly),
design takes place at the level of artifacts or systems, and components and
devices are made for that purpose, or drawn from suppliers, as the case
may be. Development work and design can occur with the suppliers, but
will depend on specifications from the customer. Network relationships
develop, but these are concerned primarily with production and sales, not
with a division of labour in design work. In principle, cosmopolitan-
ization can occur when suppliers supply to several customers and can
optimize their products according to generalized specifications. S6
Technical and other sciences can then be mobilized. The petrochemical
250 ARIE RIP
industry, for example, produces chemieals as building blocks for products

that other firms make, and does chemieal research drawing on general
chemistry, rather than developing a specialty by itself (as does happen for
their own process technology, cf. the example of heterogeneous catalysis
mentioned above). A soon as this happens, general chemieal research
becomes relevant to design work. One could call this the "design supply"
role of cosmopolitan research, in addition to its "meta-modeling" role.
The initiative at the supply side can take other forms as weil:
components or rough versions of deviees can be discovered, and
possibilities to use them in artifacts or systems are exploited. This can
happen independently from academic scientifie research. As I noted
already, the zipper is an interesting example of a device, that has set into
motion a whole series of innovations in cloth ing and coverings. The use of
X-rays in medieal apparatus, also mentioned in Section 2, is an
intermediate case: the phenomenon was discovered in a scientifie
laboratory, and in the course of basic scientific research, but it might also
have occurred elsewhere (say, in a photographer's studio), and the design
work did not draw on scientific insights (and could not do so, because
there was no systematie knowledge of X rays in the beginning). The two-
branched model presented in Section 2 can be developed further by
recognizing that exploitation of the finding is structured by the technieal
hierarchy, while exploration has a double dynamie: rationalization in
terms of devices, artefacts and systems that are being exploited, and
linking up with fundamental scientific disciplines and their theories and
accumulated empirieal insights.
By now, there are many cases where (basic or strategie) science is
working toward innovative components or deviees systematieally, for
example in polymer science and solid-state science (new materials) and in
miero-mechanies (sensors and actuators) and mechatronies . It is clear
from the latter examples that technieal-scientific fields have emerged, that
are institutionalized as fields, also in universities, and playa recognized
global role in technieal innovation. Before this could happen, however, it
was necessary that the supply of components and deviees became an
innovative activity, and recognized as something worth pursuing by itself.
The rise of the industriallaboratory in the interbellum, discussed above in
relation to meta-modeling, was also related to the interest in research into
components and devices in general, Philips Company's NatLab again
being a clear example. Subsequently, World War II efforts were a further
push: scientific knowledge and scientific manpower were mobilized on a
large scale, and whole new areas of research were created in which
fundamental work occurred, but within an overall framework defined as
component or deviee search. Polymer science, radar and other electronies
areas, and atomie energy research, are well-known examples. New
speciaIties could emerge, and the training of such specialists was seen as a
responsibility of the universities (as the pressure - but only by the I960s
- for chairs in polymer science and heterogeneous catalysis shows).
By the 1980s, three strands have to be woven into the fabrie of the
cosmopolitan design network . First, the meta-modeling role of the
technieal sciences in academic settings remains important, and links up
with the mathematieal modeling approaches that have become more
sophistieated. Second, the design supply role has become more
important, and has extended to more sciences. One element of the recent
programs to promote strategie or innovation-oriented science is exactly
the mobilization of academie sciences for design supply, with the
programs on new materials (polymers, membranes, ceramies) as key
examples. Third, innovation competition between firms and strategie
positioning of states and blocks of states (e.g. in the TRIAD) has added
a new element to technieal innovation: not the actual outcome of the
innovation is the primary aim, but the coverage of a potentially important
area of innovation. One indieator is that firms are prepared to spend (a lot
of) money on the support of academie research in order to have a
"window on science" . The coupling with local design practiees is
becoming remote, also in firms themselves. Or, aIternatively, local design
work is shaped by the exigencies of the innovation race, as can be seen
cIearly in the unending quest for the "next generation" of chips, the very
large scale integrated circuits now containing more than a million
components. Expectations always shape technologieal development
(Section 3); there is now a quasi-autonomous dynamie of strategie
research, in whieh universities can take part (fully and often only very
partially, given the complexity and resource requirements of strategie
research in many areas).
5. QUALITATIVE TRANSFORMATIONS OF THE SCIENCE-

TECHNOLOGY COMPLEX
If one thing has become cIear from the analysis in the preceding sections,
it is that there is a variety of relations between sciences and technologies,
252 ARIE RIP
that these relations and the patterns in them evolve, and that the evolution
is shaped by the socialloci of the search processes and the global scientific
and technologieal fields, with an important mediating role of the division
of labour in global design networks. By now , science and technology are
intertwined and form a complex - dancing partners that cannot be pried
apart.
The science-technology complex has been built up over decades, and its
shape shows the traces of its origins. I have focused on meso-level
developments , talking about institutions, technical domains and scientific
disciplines. These are, of course, embedded in overall societal
developments, and one should, ideally, write the history also at that level.
One attempt to do so is Tom Hughes's American Genesis, which
analyzes how the second industrial revolution is characterized by the way
the Americans built a technologieal world. This has created a
technologieal momentum: the systems, and their preferred character-
isties, take on an autonomous character and are difficult to direct or
control other than along an inertial projection. Since these mature
systems experience most of their social shaping in their early stages, they
bring out of their past, the solutions to past problems. Thus, the
momentum of the modern may be so great in the United States that the
next great technological and cultural change may occur among other
peoples in another nation."
The diagnosis of what happened is impressive, but there is a curious
neglect of the changes in science and technology that may be happening
now, in the Uni ted States as weIl as elsewhere. What is the new dance that
the dancing partners are developing? There are good reasons to think that
qualitative transformations are occurring.
One way to analyze the transformations is to conceive of an emerging
new techno-economie paradigm, as Freeman, Perez and others do. They
relate the transformation to the combination of micro-electronics and
telecommunieations becoming a so-called pervasive technology, having
repercussions in all sectors of society. 58 In doing so, their interest (as
economists) is primarily in the performances actually realized, not in the
search processes and strategie behavior that goes into the creation of these
performances. 59 Without pronouncing on the value of their conceptions,
I limit myself here to what is directly relevant to my theme of relations
between science and technology: the micro- and meso-levels of search
processes, the practices in which these are embedded, and the structures
that are emerging.
Some indieators of transformations at the cognitive side are:

• The emergence of combined technologies ("technological fusion" as it
is called by the Japanese)," whieh are recognized as such and
thematized in research programs, institutes, and perhaps, in time, new
sciences. A prime example is mechatronies, and optelectronies is an
interesting recent development. The Japanese have made a partieular
point about these combined fields: this is where the future is, they
declare, and invest accordingly . The USA National Science Foundation
program of Engineering Research Centers appears to focus on
combined technologies as weIl.
• Technologieal development leaves hardware, and becomes "paper"
(and in general, software) technology ." Within science, similar trends
are visible: Monte-Carlo calculations in physics, computer synthesis in
chemistry . Neural networks, the latest fashion , are often simulated on
computers, rather than actually built and tested . (Note how difficult it
is to decide whether neural networks are part of cognitive science or
cognitive technology.) A striking further development is how the
computer-aided design of chips (very large scale integrated circuits) is
now linked up with computer simulations of chips production
technology. Instead of real-world tests, the electronie designs are now
coupled to an electronically simulated "silicon foundry". One need not
leave the simulations any more.:" These developments create flexibility
in research, and a loosening of the links with actual technical-industrial
practiee . Models and simulations are increasingly used by industry, and
because they are less specific than concrete designs and prototypes, they
are less sensitive, and can be published in the scientific-technologieal
literature. And this is in fact what happens : for the R & D stimulation
programs of the European Community, there is little difference in
publication activity between industrial and academic participants."
• Technology becomes strategic, not in the common sense that its
products are of strategie value to actors, but that its process can be
characterized as strategie. Choices are explored rat her than technical
effects produced. This is a continuation of the previous point: in a
simulation, one can vary parameters, and find out what happens "on
paper" (paper includes electronies here). And part of the effort in
creating model systems is to explore their behavior, somewhat
independent of the extent to whieh the model systems reflect some
actual or potential technieal reality. Design is becoming an autonomous
activity??
254 ARIE RIP
The new institutional set-up contains industrial research labs, strategie

alliances, international R & 0 programs, and new hybrid centers like the
USA Engineering Research Centers . (Practitioners and their patterns of
"circulation" through the institutions are another important aspect). Let
me try to sketch the outlines of what the new institutional patterns and
structures may be." At the micro-level, a fruitful focus is to look at the
rewards that move scientists (rewards include monetary resources, but
then for the opportunities these provide , rat her than as private income).
Publishing is no longer the only criterion in academie careers. Institutions
find the amount of grants/external money that Professor X brings in
important, materially as weIl as symbolicaIly, and take that into account
in job interviews and career decisions. At first, this implies only that local,
institutional criteria will become important. If the phenomenon is
widespread however, this implies a secular change where the norms of
science become more like PLACE (instead of CUDOS), with the products of
science being defined as proprietary, i.e. related to employers or clients."
When the new externallinkages become institutionalized, however, and
industry and other sectors linking up with fundamental science adapt to
the situation, for example by allowing some exchange among scientists,
and attendant reputation building, a "new cosmopolitanism" emerges
that transcends partieular local criteria . Reputation and status can then be
acquired at the level of the field, similarly to the way reputation and status
in traditionally organized scientific fields can only be built up at the field
level if there is sufficient exchange and mobility .
This is not idle sociologieal speculation: in the case of professional
engineering, for example, engineering firms and their members often have
high professional status in spite of the reduced public character of their
work . In these engineering fields, there is enough personal mobility,
exchange at meetings, and comparisons of performance (including
promised performance, when tenders for a project are submitted and
evaluated) to make reputation and status building possible. This implies
that engineers can activelywork to acquire such reputation, because it lends
them engineering credibility that they can exploit, e.g. in further resource
mobilization. In fact, one can even find further similarities with tradition-
ally organized scientifie fields, because there will be some "organized
scepticism": the quality of products is evaluated virtuaIly, by competent
colleagues, before clients have to use the products and take real risks.
New organizations, institutions and relationships allow the emergence
of non-traditional reward systems, while their establishment allows some
of the new institutions to be integrated in the R & 0 system. To mention

a few examples: university-industry centers in the USA may evolve from
only local institutional differentiations to loci where reputation and status
at the field level can be acquired; government R & 0 pro grams become a
permanent feature of the R & 0 system, and while filled up with changing
strategie fields and priorities, participation in such programs in general
will count as important in the curriculum vitae of a researcher.
Engineering firms have been mentioned already, and more of such
"expertise brokers" appear: private medieal institutions were not
unknown, but occur more widely, and consultancy firms proliferate
(interestingly enough, also in the environment al area)."
The loci for acquiring reputation and status in science at the field level
used to be the universities, at least, after 1870. But as soon as the notion
of site for cosmopolitanism is taken seriously, there is no principle reason
for such a function to be limited to universities. Not only because
scientifie research is (and has always been) broader than university
research, and the central role of universities in the production of scientifie
knowledge may thus be a historieal accident. Also because the "strategie"
science that has become so prominent a category nowadays, is really a
label for sites and opportunities for coordinating strategie action .
Therefore, a variety of institut ions can compete for a central place. The
proliferation of "centers" connected to universities, of para-university
institutes of different kinds, and of institutes fully outside the university
system, like the Wissenschaftszentrum in Berlin, becomes a significant
phenomenon.
In other words, while it may seem that universities are only creating
new linkages (e.g. with firms) for themselves, they are in fact competing
with the other institutes on the new "market" of strategie science.
National laboratories, for example, are also becoming more active and
may weil take up new positions, depending on their ties to government,
but also on the opportunities that the changing R & 0 system offers . The
specific advantage of universities may still be that they also do (advanced)
training - but even there, competition may come up (even if earlier ideas
in this direction, e.g. scientifie training in industrial firms, or through a
consortium of firms, have come to nothing).
While these considerations are, of necessity, speculative, they do show
that the categories of "science" and "technology", or of the university as
the "horne of the scientists", 68 are of little value as entrance points to
understand the transformations that occur. Instead, one has to look at the
256 ARIE RIP
combined cognitive and socio-institutional changes. Indieative is the

struggle to create new categories to capture emerging phenomena:
" technologie de base" or "strategie science" .69 The metaphor of dancing
partners loses its point.
6 . CONCLUDING REMARKS
How far have I come in analyzing the dance of science and technology? In
two ways, I tried to overcome barriers to our understanding: I did not
start with science, its nature and its possibilities with regard to technology,
but as it were turned the tables, by looking at search processes in general,
and at technology being rationalized, becoming theoretieal, " scientific"
even, and thus offering opportunities for existing science to get involved.
Secondly, I emphasized the danger of following legitimatory usage of
science and technology as umbrella terms," and focused on ongoing
processes and institutionalized activities instead. This is not to say that
there is no such thing as "science" , or "technology", or that it makes no
sense to talk about the world of science and the world of technology, and
how these interact. But the notion of science as such is constructed (and
maintained) by actors in specific contexts, for example when in the 19th
century calls are made for engineering to become scientifie, or when
science is defended as being the backbone of Western civilization. The
force of science does not exist in a vacuum , but derives from these
contexts and what actors do with such constructions. SimiIarly, there are
times and piaces where worlds of science are created, and become
sufficiently institutionalized to continue as such. Engineers (and other
people not usually seen as scientists or scholars in the 19th century) can
then be said to follow hybrid careers when they also playa role in the
world of science." For the late 20th century, however, it may be
misleading to analyze in terms of a world of science, that is a direct
successor to the world of science as it existed, at least for physicists in the
I920s. 72 That is why I have emphasized that transformations occur: while
there is some continuity of actors, institutions and practiees, the overall
pattern shifts, and it now means something different to be in the world of
science.
In this approach, there is the unavoidable ambiguity of using common
sense notions of scientists and engineers, of science and technology and
their demarcations, to mobilize empirieal material and interpret it, and on
the other hand not wanting to be imprisoned by such categories which will
hide rather than disclose. I have taken a sociological detour through
search processes, heuristics and expectations, local design work and
cosmopolitan design networks with their division of labour, to be able to
reconceptualize what is happening in the science-technology complex.
Central to this reconceptualization is the relation between local search ,
design and anticipation activities, and repertoires, cognitive-technical
infrastructures and interactions and institutions at the field level. In these
terms, one can give an analytic definition of what science is, if one wishes
to do so: when search products are "consumed", that is, assessed and
used, by others like yourself, and a cognitive and evaluative repertoire has
emerged together with sociallinkages and mobility implying community
formation, it makes sense to speak of (a) science. In contrast, technology
has search processes which are directly or indirectly related to local design
and construction activities . (Note that this definition makes large parts of
social science and policy analysis fall under technology.) The dancing
partnership of science and technology now becomes a relation between
activities oriented to different reference points and groups, rat her than a
matter of combining different cognitive-technical repertoires.
Obviously, in concrete cases people draw on different domains and
specialties, some of which have dynamics like science, and others like
technology. At this level, there may weIl be problems of interaction, but
these relate to the general problem of combining insights and skills from
different sources . Similarly, institutional problems, like interaction
between university and industry, or new roles for academic research, need
not derive from any general issue of "science" versus "technology", but
from heterogeneity of institutions. As I have argued, such heterogeneous
social loci will, in fact , feed into new definitions and demarcations of
science and technology.
In a sense, I define the question of the relation between science and
technology away . In so far as it is to be discussed, it is in a battle of
legitimat ions . To analyze and understand interactions, one had better not
use the umbrella terms.
Looking back at the alternative analyses and conceptualizations I have
discussed in this paper, however, these do not completely exhaust the
original question. One can do away with "science" and " technology" ,
because these are rhetorical constructs rat her than actual dancing
partners, but one still has to face the question, at the micro-Ievel, of the
relation between knowledge and artifacts .
258 ARIE RIP
To bring out a way to approach this question, the difference between

knowledge and artifact has to be reconceptualized . The received view runs
as folIows:
Although science and technology both involve cogn itive processes, thei r end result s are not
the same. The final product of innovative scientific activity is most Iikely a written
statement, the scientific paper, announcing an experimental finding or a new theoretical
position. By contrast, the final product of innovative technological activity is typically an
addition to the made world: a stone hammer, a clock , an electric motor.?
The contrast, however, cannot be made so easily.

Implicit in my analysis is the idea that there is a continuum rather than
a categorical difference. What is used is not knowledge as such, but
knowledge claims: statements which include the claim that the insight or
effect is applicable in other situations, including the situation at hand.
While artifacts, at least as they are treated in technology, are not just
contraptions laying around, but a promise of functionality provided one
embeds and uses the artifact in the right way. Placing knowledge claims
and artifacts on a continuum in this way allows me to argue that both can
be inserted in search processes, without any special problem of a
categorical divide (compare the beginning of Section 3). The two-
branched model of exploitation and exploration of a finding, discussed in
Section 2, in fact builds on this insight; insofar as different routes are
followed, this is a matter of division of labour and institutionalization,
rat her than a difference in kind .
The interesting difference relates to the decontextualization and
cosmopolitanization that happened to knowledge claims and to design
and construction insights and rules. Decontextualization/cosmopolitan-
ization creates a certain robustness, but also distance to concrete, local
usage. It is this tension , and the attempts to resolve it, which creates the
movement that I have described in this paper. Various cross-sections and
other constructions can be labeled "science", and in other cases
"technology" . The dance of "science" and "technology" is then like the
wooden figures on top of a musical box: they turn and bow gracefully, but
are puppets rat her than actors.
University 0/ Twente
NOTES
1 In Price [l984a) he adds that what keeps them linked is that both dance to the music of
instrumentalities; see also Price [I 984b).

, Latour & Woolgar (1979) ; also Rip [1982) .
J Cf. Latour's (1987) criticism of the diffusion model. In Latour (1990) he iIlustrates the
necessity of working on the situations with a number of exarnples , including one from a
study by Marc Bloch: " In the late Middle Ages, the grinding stones, the gears, the wheels
and the rivers are good unexpected allies that , once tied together in one mill, makes a
formidable stronghold. But their efficiency stops there. A stronghold can be in the middle
of a battlefield , thus bearing on the issue of the battle , or away from the battlefield . If each
household goes on grinding corn by hand , the Prince, who holds the communal mill, will
hold not hing but wood , water and stones. The mill will become a stronghold only if the
Prince fetches the militia, enforces the King's ruling, the Church 's teaching and compels
every household to break their hand-grinders and to pass through the miller's stone ." (p. 26)
The moral of this story , as Latour makes c1ear, is tha t social power is necessary to ensure
technical power (as well as vice versa), In his harangues against technological determinism
(labeled " trajectories" in the 1990paper) he gets carried away by his own rhetoric, however,
and forgets that there may be patterns in the sociotechnical alliances that might usefully be
described as trajectories .
• See Wolfgang Krohn's introduction to Zilsel (1976), esp. pp . 23-29 .
, Homburg (1983).
• Casimir (1983) . Latour and Woolgar (1979).
, Rip (1985).
• Böhme, Van den Daele and Krohn (1978) .
• This idea draws on Disco, Rip, Van der Meulen (1992) and ongoing work of these authors.
See further Section 4.
10 Nelson and Winter (1977) , Dosi (1982) , Dosi et al.. (1988).
11 Quoted after Van den Daele [19781 , p. 32.
" See for example the way even Dosi (1982) essentially works with a sequence model, where
findings go downstream from science to technology, passing filters (that include socio-
economic criteria) .
13 Ziman [1976) , pp . 180-181.
" Ziman (1976) , pp. 188-193.

" As described by McMiIlan (1979) .
16 McMiIlan (1979) , p. 89.
I.
" McMiIlan [1979), pp. 90-91.
For our argument here, the starting finding may be treated as a point source . When one
looks more c1osely, it dissolves into contingent processes, and only with the benefit of
retrospection can one identify a source. Compare Latour's (1987) analysis of the Diesel
engine.
" McMiIlan (1979).
'0 Prakke [1988].
" In a study on knowledge transfer (Rip, Schoemaker and Meus [1983]), it was argued that
the "matching' of university research and societal needs requires cognitive and social
transformations, both the aggregation of research questions through sectoral organizations,
260 ARIE RIP
patients ' associations, trade unions ; and disaggregation of academic approaches, and that
one can observe such processes in actual successful knowledge transfer. Note further that
laboratory-based, in the main text, should be read as including method-based (e.g. social
science methods like surveys). The point is that scientific findings derive their "universality"
from being produced in restricted environments (see Rip (1982)); this point is the same as
Harre's, when he argues that "manipulation of matter under the control of a skill is prior
to a more conventional scientific activity' (Harre [1990], p. 35). In the complexities of the
real world, applicability is not assured by "universality", and requires either de-
universalization or concretisation (Böhme, Van den Daele and Krohn (1973)) or
transforming the world into something resembling the laboratory (Rip [1982], Latour
(1983)).
22 This is also argued by Latour, but then from apower perspective, when he focusses on the
Machiavellian engineering that goes on in all walks of life (Latour (1986), (1987)). The
search perspective is not blind to power, but takes the cognitive aspect seriously. In the
search perspective, there is the same attention to the need to find and enroll allies, both
human and non-human: microbes and electrons have to be kept under control in order to
have robust findings . But there is also attention to exploration of possibilities, following a
lead, elaborating a cognitive schema . This is not sufficient to explain the "success", i.e. the
robu stness of findings. But it is necessary to understand the direction that is taken .
2J This view of scienceand technology has been put forward by a number of authors by now,
including de Solla Price and Layton, and has been dubbed the "two cultures model" by
Barnes and Edge [1982, pp . 147-154, esp. p, 151]. See also Weingart [1978] for a
differentiation-scientification thesis which is quite compatible with the present anal ysis.
24 The term "bricolage" , in the sense used here, has been introduced by Levy-Strauss, who
distinguished between "esprit d'Ingenieur" (that is, the French graduates of the Grandes
Ecoles with their mathematical bent), and the "esprit de bricolage" (quoted after Barnes
(1974), p. 58, 146). I adapt the term further to describe those aspects of science and
technology where the researcher scouts for whatever he can use, and is prepared to make do
with what he finds .
25 Nelson and Winter's [1977, 1982) argument focusses on "routines" and " routines to
develop new routines " (in their terminology), where routines stand for ways of doing things,
while rout ines to develop new routines are heuristics of search processes.
2. Expectations about new scientific findings or technological possibilities are voiced, and
acted upon , continually. The early promise of biotechnology has attracted biologists, R&D
firms, and venture companies . The discovery of superconductivity in materials at liquid
nitrogen temperatures had a magnetic effect on physicists and chemists, as weil as funding
agencies all over the world. Expectations, however, are not limited to such highly visible
examples of early prom ises. They are at play already when a researcher chooses between
options to folIowand does so in terms of the expectation of success.
27 The notion was introduced by Van den Belt and Rip (in Bijker et al. [1987], p. 140: " ...
there have to be expectations ab out the success of continuing work within [a particular]
cluster of heuristics - expectations that must be embedded in the subculture of the technical
practitioners and others involved in the development." They add in a footnote: " Sometimes
the relevant expectations may be of a broad, even sociological nature . In the case of the
French VEL project, the EDF engineers working on this particular project acted on an
explicit vision of where French society was going - avision that was characterized by Callon
as a variant of Tourainian sociology . One can also think of the Japanese Fifth Generation
project, which seems to have been predicated on, inter alia , Daniel BeIl's ideas on the coming
of the post-industrial society. UnIike Touraine for the French EDF engineers, however, Bell
became an acknowledged ideologue for the Japanese "knowledge engineers [...[." See for
more extended discussion and empirical cases of the cultural matrix of expectations,
Vergragt, Mulder, Rip en Van Lente (1990).
,. Van den Belt and Rip [1984, 1987).
,. By contrast, the structural formula proposed by Kekule in 1866 (using a theory of organic
chemistry which was more correct according to our present day view) mere!y amounted to
the translation of Hofmann's formula into his own theory but did not indicate new
directions for innovation. Instead of immediately leading to technological applications, the
contribution of Kekule 's structure theory must be found in the way it affected the cultural
matrix of synthetic dye chemistry. It was only in 1880 that the constitution of rosaniline was
finally elucidated by Emil and Otto Fischer in terms of Kekule's theory. This elucidation was
the source ofa new sequenceof dye innovations. (Quoted from Van den Belt and Rip (1987),
p .I44.)
30 Van den Belt & Rip (1984) note that the Muttersubstanz heuristic was after a time used in
new areas of appIication, especially pharmaceuticals, and became a heuristic in synthetic

chem istry in general. This shows how heuristics can be adopted more widely than for the
technological deve!opment in/through which they occurred orig inaIly.
31 Here, our terminology is the same as Dosi 's (1982). Thi s concept of "trajectory" should
be distinguished from Nelson and Winter's "natural trajectories", which are, in our
terminology, stabilized heuristics. See their (1977), p, 56: "[ ... ] there are certain powerful
intra project heur istics that apply when a technology is advanced in a certai n direct ion, and
payoffs from advancing in that direction that exist under a wide range of demand
conditions. We call these directions "natural trajectories" . If natural trajectories exist,
following these may be a good strategy." Nelson and Winter's equ ivalent of our
technological paradigm is " technological regime" ,
32 Note that there is no impIication that technological developments always follow
trajectories. A trajectory is a contingent happening . This point underscores my difference
with Latour (compare notes 3 and 22). " Trajectory" does not imply any technological
determinism . It indicates that search processes are sometimes oriented in specific ways,
33 Constant [1984].
3. Homburg (1983).
35 This was recognized as such at the time, and led to problems with the patent law, which at
that time required inventiveness for a patent to be awarded , so did not allow products from
routinized research to be patented. The solution was to add an escape clause: if a commercial-
ly useful effect was found , the finding was still patentable (Van den Belt and Rip (1987)).
3. This observation is related to the debate, started by Latour [1987], whether one should
include non -human as weIl as human actors in the network dynamics. My position is that to
understand technological dynarnics, one should include expectations and the way these
anticipate and mobilize . Voicing expectations , and being swayed by them, are Iimited to
human actors. In fact , Callon (in Callon et al. [1986] and Callon (1987)), in his notion of
actor-world as a scenario created and pushed by a human or institutional actor (in his case,
Electricite de France) is using this notion already. Electrons can act, are actors (or agents) in
that sense , but cannot interact on the level of anticipations and legitimations,
262 ARIE RIP
Actor-network theory focusses on processes, as if legitimations were irrelevant. This is a

useful antidote to the exclusive attention to the cognitive level, but neglects the search and
expectation part of technological dynamics. One could probably describe technological
developments over time with curves for the (related) developments on the process/network
level, and on the cognitive/legitimation level, and indicate where the one was dominating the
other. It would then become an empirical quest ion how important the two aspects of
technological dynamics are .
3 7 "When a core conc ept has emerged, it becomes the top of a design hierarchy.
Standardization of product design changes the basis of competition . Battles in the market
place are no longer fought over the kind of thing a product is or even the kinds of things it
should be able to do . The locus of competition shifts to what the product costs . (... ) This is
not to say that product innovation disappears ent irely or that it ceases to be of value
aItogether. The point, rat her , is that what product innovation there is tends to be localized
toward the bottom of established design hierarchies and, as a consequence, to enjoy little
market visibiIity' (Abernathy er al. (1983). p. 24). See also Clark (1985).
38 Dosi (1984). pp . 93-94. 194-195 .
3. A similar observation can be made about the Starnberg model of development of sciences
which has been used to der ive policy impl ications, for example the differential possibilities
of orientation towards social relevance in different phases of development (Schäfer (1983).
Rip (1981)).
40 Van den BeIt and Rip (1987); see also Schot (1991).
4. Genentech, the first of such firms to offer its shares on the open market, was introduced
to the New York Stock Exchange in October 1980. Its issue price was $35. but prices jumped
to $89 in just one hour, to stabilize at $75 by the end of the first day . This phenomenon has
been quoted often, while much less notice has been taken of the fact most of the increase
disappeared Inter on: by December 1980. Genentech stock stood at $45.
42 The field level is less visible in another locus for negotiation of expectations: inside firms.
It is difficuIt to make general pronouncements, because not many cases have been
documented yet. One example is the way the strong fibre Twaron was developed within
Akzo Company. At one moment, the board of directors decided to stop the developrnent ,
but the network of alliances and expectations lower in the firm had become so strong that
work continued (and ach ieved success in the end) . See Mulder and Vergragt (1988).
43 Cf. Bijker, Hughes and Pinch (1987).
44 The example of engineering design will indicate the possibilities of this approach. A
design or a blueprint is not just a set of instructions, a virtual artefact as it were, but also a
story ab out how functions can be fulfilled and specifications can be met. The story must be
convincing to move others into realizing the artefact (or an artefact more or less like the
design intended); in addition, chances for its eventual survival should be increased. Thus, a
good design has to take further requirements into account: the mobilization of varied
resources for its realization and survival. It is a matter of heterogeneous engineering to
develop working technology. Law (1987)
Engineering design is also the place where one should anticipate on future niches and future
expansion. If all this is taken up in the "story telling", technological development would
become reflexive. i.e . self-conscious about its dynamics. In practice, however , there are
severe limitations. For one thing, there is too much division of labour, which induces
modular optimization . Computer-aided design, an important tool , helps to experiment with
meeting specifications, but lets designers forget ab out heterogeneous engineering . The
danger is that so-called path-dependencies, unavoidable in quasi-evolutionary develop-
ments, will become counter-productive. Soete [1988J, p. 56, quotes Arthur and David for an
example of locked-in technological development due to path dependency: the QWERTY
keyboard of typewriters and almost all keyboards, including those of computers, which is
not the most efficient outlay, but which has resisted attempts at change . The introduction,
in the late 19th century, of this kind of outlay of letters was related to the requirement of
slowing down the speed of the typist. to avoid the hammers getting cluttered all the time. See
David [1985). In other words, rationalization of design should not just be optimalization of
particular bits and pieces, but take process rationalization as starting point.
•• This section draws heavily on Disco, Rip and Van der Meulen [l992J.
•• Compare Vincenti's discussion of normal design in relation to normal configuration of a
device, e.g. in Vincenti [1991J, and Vincenti, this volume.
•, Pambour's theory of the steam engine is a good example, also because it was widely used
to improve local design work (see Kroes, this volume). The attempts, for example by
Reuleaux in the 1860sand 1870s, to formulate a general theory of machines are almost too
abstract. In mechanical engineering curricula, however, one sees definite indications of such
a generalized approach .
•• Mathematization of design had already been prepared in France in the circles of the Corps
des Ponts et Chaussees and among Polytechniciens in general, for example in the elaborate
theory of suspension bridges published by Henri Claude Navier in 1823(and which won hirn
admission to the Academie in the following year) or the thermodynamic theory of steam
engines published by Sadi Carnot in the 1820s. For Navier and the social origins of French
theoretical mathematical modelling see Kranakis [1989J . Carnot's work itself is not
mathematical; later Clapeyron and Clausius elaborated actual mathematization of
thermodynamics.
•• A similar point can be made (and has been made) about the importance of teaching
practices, and the teaching function in general , for the emergence of theory in science, e.g.
by Janich [1978), and is prefigured in Böhme et 01. [l978J analysis of medieval cathedral
building. In Rip [1982J, I add to this argument by looking at technical dynamics (i.e. control
of conditions and effects) within science, and show that "Increasing restrictedness leads to
empirical generalizations and conceptual distinctions, i.e. "bottorn-up" theory formation .'
(p. 231) Generalization is important in science to spread one's know1edge claims more
widely; this becomes important in technology as soon as technical sciences have created a
domain for themselves, with colleague technical scientists rather than designers and
practitioners as "consumers' . The point in the main text is that generalization is already
stimulated by the need to spread one's trainees more widely.
s 0 In radio engineering, on the other hand, the utter novelty of the technology implied a
diffuse and uncertain theoretical base and the consequent dominance of practice over
theory . Technical models tended to be intuitive or at best highly speculative and, such as they
were, to be generated mainly within firms manufacturing wireless equipment and within
state agencies with mandates in this domain. Only after radio technology began to stabilize
in the mid-I920s did engineering schools become significant research sites for "meta-
modelling" (generally aimed at high-level theoretical models of waveforms, propagation,
etc.) and did reasonably elaborated cosmopolitan design networks emerge. Cf. Aitken
[1978J; Disco [199OJ discusses the cases of radio engineering and reinforced concrete .
264 ARIE RIP
51 Buchholz [1979]; Rip and Van der Es [1980]. Also Vermeij [1990] for a glimpse of the
local design work in the biotechnology of waste treatment.

" See Freeman [1974], for the history of Catalytic Research Associates , a consortium of the
big companies that developed catalytic cracking of oil in the 1930s and early 1940s. While
the rneta-modelling of such technologies had gone a long way in industry, it was only in the
1960s that professors (at first , part-time) were appointed. Th is is probably related to a
tradition of black boxing of the catalytic process (as a component or device, cf. discussion
below) in chemical technology and process technology, and a renewed interest when research
at the microscopic level became possible .
53 Medicine might appear to be an exception but was, at the time, very little scientificized
and instrumentalized, and in so far as it did work as a technology, could be ideologically

papered over as really belonging to the realm of therapy.
The situation in the US is complicated because the traditional universities, involved in a long
struggle to emancipate themselves from an ideology of teaching and doing practical work,
did not want to fall under the newly created label of "applied science" (cf. Daniels [1967]),
and attempted to imitate the German research university - while at the end of the century,
land-gram colleges (including the Universit y of California and other, by now well-known
universities) could be established with tax relief etc . provided they worked in the public
interest.
54 It should be stressed that thi s academic distance from emergent cosmopolitan design
networks was typical of the European continent and that other developmental patterns
prevailed elsewhere. In Britain and especially the United States, where engineering schools
were often tacked on to universities and where universities in any case had long experience
in interfacing with practical fields (e.g. agriculture, industry, and mining) academic research
developed an applied orientation early on in the game .
es The not ion of a design hierarchy is used to structure the division of labour in specific
design tasks, e.g. of an airplane (see Vincenti [1991], and this volume) . It has also been used
10 describe the overall pattern of design in an industry or sector (see Clark [1985]), where the
dominant design at the top functions like a paradigm . I introduce aseparate concept ,
technical hierarchy, to emphasize the stratified character of modern technology in general,
which allows for a generalized division and coordination of labour.
56 It is for this reason that in Pavitt's [1984] taxonomy of innovation patterns supplier-
dominated innovation comes out as one of the four main patterns. Suppliers can also take
initiatives, based on their generalized position, e.g. when in the 1960s the petrochemical
industry developed biodegradable fatty acids for use in synthetic detergents (to overcome
environmental problems) rat her than the detergent industry itself (Daey Ouwens et 01.
[1987], Ch . 4). Note that the research was of a problem-solving nature, and thus did not lead
to a lasting pattern of division of labour.
57 The quotes are from Hughes [1989], p . 2 and 471, while the description of technological
momentum is taken from Hughes [1988], p. 92.

58 The notion of techno-economic paradigm is pervasive in the chapters of Dosi et 01. [1988].
While they do not use a linear model for the relation between science and technology, they
seem to assurne such a pattern for the new electronic and communication technology, as it
spreads through society , Freeman, accordingly, has been emphasizing the need to invest in
the implementation of these technologies.
59 Rip [1992].
6. Kodama (1990).
6' Laredo and Callon (1990), pp. 166-167 .
62 Rip [I989b) refers to IEEE Spectrum [Oct . 1988), pp , 24-28, for the simulation link
between design and manufacturing of VLSI. He also suggests that the innovation race to
create ever larger chips could now be done completely on computers: war games instead of
actual war. Orson Scott Card's science fietion novel Ender's Game offers an interesting (and
moving) portrayal of the ambiguities of such a situation.
6J Laredo and Callon (1990).
6' There is (I think) complete equivalence with scenario building to explore futures and
articulate decision making. This would explain the similarity with IIASA type model
building (cf. energy, acid rain) and other mixed technical-social modeling, e.g . of transport
systems , The logistics of transport systems (including their embedding) become object of
technologieal development work .This is clear, for example, from the evolution of the
relevant departments in Delft Technologieal University. See P rins (1987).
" See Rip (1990) and (1991) for a more extensive discussion.
" The acronym PLACE was coined by John Ziman to indieate the Proprietary, Local,
Authoritarian, Commissioned and Expert nature of industrial research and a lot of present-
day scient ifie research, and contrast this with the Mertonian "CUDOS" norms:
Communalism, Un iversalism, Disinterestedness, Organized Scepticism. (Ziman (1983),
(1990)).
67 Professional engineering has evolved its own kind of " scientizing" over aperiod of a
century or so ; a similar argument can be made for more recent developments, e.g . in expert
advice on health, on environment and on global issues. Then, it is understanding relevant to
decision-rnaking that counts. So me (and sometimes most) of the work is not made public,
and if public, often in reports rather than "regular' scientifie publi cations (compare the
increasing prominence of grey literature). Still , there is recognition of performance, and
accompanying rewards in terms of resources and careers. So scientists can (and will) move
in this direction. Clearly, the traditionally organized R&D system is not sacrosanct, in an y
case not something for whieh many scienti sts will sacrifice career chances, and actually
possibilities of research. And one can in fact see some "new cosrnopolitanism" , becau se
hybrid institut ions emerge like mixed scientifi c-policy conferences, organizations like the
International Institute of Applied System s Analys is near Vienna that create career resources
and mobility. In this way, contributions of experts will still be visible and a fun ctioning
reward system can evolve again .
6. The term has been used , see the subtitle of Wittrock and Elzinga (1985), but without the
critieal assessment required.
6. For "technologie de base " see Laredo and Callon (1990), for " strategic science " Irvine
and Martin (1984), and Van Lente and Rip (1991). In fact , the word "technology" itself used
to be such a new category: At the Chicago World Fair of 1933, the linear model ideology was
still expressed as "Science Finds, Industry Applies, Man Conforrns" (my emphasis) . Use of
the term " technology" is related to legitimation strategies of engineers and to the
transformations during and after World Wa r II .
7. Cf. also Rip (1989).
11 Kranakis, this volume.
72 This world is described by Casimir (1983) and McCormmach (1983).
73 Basalla [1988], p. 30, as quoted in Gremmen (1991), p. xiii.

266 ARIE RIP
REFERENCES
Abernathy, William J ., Clark, Kim B., and Kantrow, Alan M.: 1983, Industrial Renaissance
- Producing a Competitive Future for America, New York : Basic Books.
Aitken, Hugh G .J .: 1978, 'Science, Technology, and Economics: The Invention ofRadio as
a Case Study,' in Krohn, Wolfgang, Layton, Edwin T . and Weingart, Peter (eds.) The
Dynamics 0/ Science and Technology, pp. 89-112.
Barnes, Barry: 1974, Scientific Knowledge and Sociological Theory , London: Routledge
and Kegan Paul.
Bames, Barry, and Edge, David (eds.): 1982, Science in Context, Milton Keynes: Open
University Press.
Bijker, Wiebe E., Hughes, Thomas P. , and Pinch, Trevor 1. (eds.): 1987, The Social
Construction 0/ Technological Systems. New Directions in the Sociology and History 0/
Technology, Cambridge, MA : MIT Press.
Böhme, Gernot , van den Daele, Wolfgang, and Krohn , Wolfgang: 1973, ' Die Finalisierung
der Wissenschaft' , Z. f Soziologie 2, 128-144.
Böhme, Gernot, van den Daele, Wolfgang, and Krohn, Wolfgang. 1978, 'T he
"Scientification" of Technology", in Krohn , Wolfgang et al. (eds .) The Dynamics 0/
Science and Technology, pp . 219-250.
Buchholz, Klaus: 1979, 'Die gezielte Förderung und Entwicklung der Biotechnologie", in
Van den Daele , Wolfgang et al. (eds.) Geplante Forschung, pp . 128-144.
Callon, Michel : 1987, 'Society in the Making: The Study of Technology as a Tool for
Sociological Analysis', in Bijker, Wiebe E . et al. (eds .) The Social Construction 0/
Technological Systems. New Direct ions in the So ciology and History of Technology, pp.
83-103 .
Callon, Michel, Law , John, and Rip, Arie: 1986, Mapping the Dynamics 0/ Science and
Technology, London: Macmillan .
Casimir, Hendrik: 1983, Haphazard Reality , A HalfCentury 0/ Science , New York: Harper
and Row .
Chubin, Daryl E . and Chu, Ellen W . (eds .) : 1989, Science 0// the Pedestal. Social
Perspectives on Science and Technology, Belmont, CA : Wadsworth Publ.
Clark, Kim B.: 1985, 'The Interaction of Design Hierarchies and Market Concepts in
Technological Evolution", Research Policy 14, pp . 235-251.
Cole, Gerald S.: 1988, 'T he Changing Relationships Between Original Equipment
Manufacturers and Their Suppliers' , Int. J. Technology Management 3, pp. 299-324.
Constant, Edward W. 11: 1984, 'Communities and Hierarchies: Structure in the Practice of
Science and Technology', in Laudan, Rachel (ed .), The Nature 0/ Technological
Knowledge. Are Models 0/ Scientific Change Relevant? Dordrecht: Reidel.
Cozzens, Susan E., Healey, Peter, Rip , Arie, and Ziman, John (eds.): 1990, The Research
System in Transition . Dordrecht: Kluwer Academic Publishers.
Daey Ouwens, Cees, van Hoogstraaten, Pieter, Jelsma, Jaap, Prakke, Frits, and Rip , Arie:
1987, Constructief Technologisch Aspectenonderzoek, Een verkenning. 's-Gravenhage:
NOTA.
Daniels, George H. : 1967, "The Pure-Science Ideal and Democratic Culture', Science 156,
pp. 1699-1705.
David, Paul A .: 1985, 'Clio and the Economics of QWERTY' , Amer. Econ . Rev, 75, pp .
332-337.
Disco, Cornelis, Rip, Arie, and van der Meulen , Barend: 1992, "Technical Innovation and the
Universities : Divisions of Labor in Cosmopolitan Technical Regimes', Social Science
Information 31, pp . 465-507.
Dosi, Giovanni: 1982, "Technological Paradigms and Technological Tra jectories: A
Suggested Interpretation of the Determinants ofTechnical Change', Research Policy 11,
pp. 147-162 .
Dosi, Giovanni, Freeman, Christopher, Nelson, Richard, Silverberg, Gerald, and Soete, Luc
(eds.): 1988, Technical Change and Economic Theo ry , London: Pinter.
Elias, Norbert, Martins, Herminio, and Whitley, Richard (eds.): 1982, Scientific
Establishmentsand Hierarch ies, Sociology of the Seiences Yearbook, Vol. VI, Dordrecht:
Reidel.
Elliot, Brian (ed.): 1988, Technology and Social Process , Edinburgh : Edinburgh University
Press .
Freeman, C.: 1974, The Economics of Industrial Innovation, Harmondsworth, Penguin .
Gremmen, Bart (ed.): 1991, The Interaction between Technology and Science , Wageningen :
Department of Applied Philosophy, Wageningen Agricultural University .
Harre, Rom : 1990, 'The Dependence of ' Hi-Tech' Science on Technology through the Role
of Experimental Technique', RISESST 1, pp. 35-42.
Homburg, Ernst: 1983, "The Influence of Demand on the Emergence of the Dye Industry.
The Roles of Chemists and Colourists', J. Soc. Dyers and Colourists 99, pp. 325-333.
Hughes, Thomas P . : 1983, Networks of Power. Electrijication in Western Societies, 1880-
1930, Baltimore, MD : Johns Hopkins University Press.
Hughes, Thomas P . : 1986, "The Seamless Web : Technology, Science, Etcetera, Etcetera',
Social Studies of Science 16, pp. 281-292 .
Hughes , Thomas P .: 1988, 'Inventing Controls for Large Technological Systems', in Mayer ,
Evelies (Hrsg .), Ordnung, Rationalisierung, Kontrolle, pp . 83-94, Darmstadt:
Technische Hochschule.
Hughes, Thomas P.: 1989, American Genesis: A Century of Invention and Technological
Enthusiasm, 1870-1970. New York, N.Y.: Viking .
Irvine, lohn and Martin, Ben R.: 1984, Foresight in Science. Pieking the Winners, London:
Frances Pinter.
Janich, Peter: 1978, ' Physics - Natural Science or Technology?' in Krohn , Wolfgang et al.
The Dynamics of Science and Technology, pp . 3-27.
Knorr-Cetina, Karin D. and Mulkay , Michael (eds.): 1983, ScienceObserved. Perspectiveson
the Social Studies of Science, London: Sage.
Kodama, Fumio: 1990, 'Japanese Innovation in Mechatronics Technology', in Sigurdson,
Jon (ed.) Measuring the Dynamics of Technological Change, pp. 39-53, London: Pinter
Publishers.
Kranakis , Eda : 1989, 'Social Determinants of Engineering Practice: A Comparative View of
France and American in the Nineteenth Century', Social Studies of Science , pp. 5-70.
Kroes, Peter: 1992, 'On the Role of Design in Engineer ing Theories; Pambour's Theory ofthe
Steam Engine ' , this volume .
Krohn, Wolfgang. Layton, Edwin T ., Jr ., and Weingart, Peter (eds.): 1978, The Dynamics
ofScienceand Technology, Sociologyofthe Seiences Yearbook, 1978, Dordrecht: Reidel.
Kuhn, Thomas S. : 1970, The Structure ofScientific Revolutions, 2nd enlarged ed ., Chicago :
University of Chicago Press.
268 ARIE RIP
Laredo, Philippe, and Callon , Michel: 1990, L 'imp act des programs communautaires sur le
tissue scientifique et technique franca ise, Paris : La Documentation Francalse.
Latour, Bruno : 1983, 'Give Me a Laboratory and I Will Raise the Wor1d' , in Knorr-Cetina
and Mulkay (eds.), Science Observed. Perspectives on the Social Studies 0/ Science, pp.
141-170.
Latour, Bruno: 1987, Science in Action , Mihon Keynes: Open University Press .
Latour, Bruno : 1988,' How to Write "The Prince" for Machines as weilas for Machinations',
in Elliot , Brian (ed.) Technology and Social Process, pp. 20-43 .
Latour, Bruno: 1990, ' The Prince of Machines (ltalian version)', RISESST I , pp. 15-34 .
Latour, Bruno , and Woo1gar, Steve: 1979, Laboratory Life. The Social Construction 0/
Scientiflc Facts, Beverly Hills and London: Sage.
Law, John: 1987, 'Techno1ogy and Heterogeneous Engineering: The Case of Portuguese
Expansion', in Bijker, Wiebe E. et al. (eds.) The Social Construction 0/ Technological
Systems. New Directions in the Sociology and H istory of Technology, pp . 111-134.
McCormmach, Russell: 1983, Night-Thoughts 0/ a Classical Physicist, New York: Avon
Books (orig. Harvard V.P . 1982).
McMilIan, Frank: 1979, The Chain Straighteners, London, Macmillan .
Mu1der, Kare! and Vergragt, Philip: 1988, 'Weaving the Super-Fibre Network' , paper, 4S-
EASST Conference , Amsterdam, 16-19 November 1988.
Nelson, Richard and Winter , Sidney: 1977, ' In Search of a Useful Theory of Innovation',
Research Policy 6, pp. 36-76.
Nelson, Richard and Winter , Sidney: 1982, An Evolutionary Theory 0/ Economic Change,
Cambridge MA: Belknap Press of Harvard University.
Pavitt, K.: 1984, 'Sectoral Patterns of Technical Change: Toward s a Taxonomy and a
Theory', Research Policy 13, pp . 343-373.
Pinch, Trevor J . and Bijker, Wiebe E.: 1984, 'The Social Construction of Facts and
Artefacts' , Social Studies 0/ Science 14, pp . 399-442 .
Prakke, F.: 1988, 'Manufacturing Innovation Between the State and the Market ', in
Gerstenfeld, A. and Warnecke , H.J . (eds.) Proceedings of Conference on Manufacturing
Research: An International Perspective, Amsterdam, Elsevier.
Price, Derek J . de Solla: 1965, 'Is Technology Historically Independent of Science? A Study
in Stat istical Historiography', Technology and Culture 6, pp . 553-568 .
Price, Derek J . de Solla: 1984a, 'Notes Towards a Philo sophy of the Science/Technology
Interaction,, in Laudan, Rache! (ed.), The Nature 0/ Technological Knowledge. Are
Models 0/ Scientific Change Relevant?, pp . 105-114, Dordrecht: Reidel.
Price, Derek J . de Solla: 1984b, 'The Science/Technology Relationship , the Craft of
Experimental Science, and Policy for the Improvement of High Technology Innovation',
Research Policy 13, pp. 3-20 .
Prins, G.: 1987, Een hele hijs. Afscheidsrede hoogleraar in de Transporttechniek, Delft:
Technische Universiteit .
Rip, Arie: 1981,'A Cognitive Approach to Science Policy', Research Policy 10, pp. 294-311 .
Rip, Arie: 1982, 'The Development of Restrictedness in the Sciences', in Elias, Norbert et al.
(eds.) Scientific Establishments and H ierarchies, pp. 219-238.
Rip, Arie: 1985, ' Interdisciplinarity without Disciplines? A View from the Sociology of
Science', in Mar, Brian W. , Newell, William T., and Saxberg, Borje O. (eds.) Managing
High Technology, pp. 185-192, Amsterdam: Elsevier Science Publishers.
Rip, Arie: 1986, 'The Mutual Dependence of Risk Research and Politiea1 Context', Science
and Technology Studies 4(3/4) (Fall/Winter), pp. 3-15.
Rip, Arie : 1989a, 'Opening the Black Box of Technological Development', in Chubin, Daryl
E. and Chu, Ellen W. (eds.) Science offthe Pedestal. Social Perspectiveson Science and
Technology, pp. 137-144 .
Rip, Arie: 1989b, 'Oe partiele rationaliteit van VLSI ontwikkeling en gebruik', T. Nedl.
Elektr. and Radio Genootsch. 54(2), pp . 55-60.
Rip, Arie: 1990, 'An Exercise in Foresight : The Research System in Transition - to What?',
in Cozzens, Susan E. et al. (eds.) The Research System in Transition, pp . 387-401.
Rip, Arie: 1991 , 'The R & 0 System in Transition: An Exercise in Foresight', in Science and
Social Priorities - Perspectives of Science Policy for the /990s, Proceedings of the
conference sponsored by International Council for Science Policy Studies, UNESCO, and
Institute of Theory and History of Science, Czechoslovak Academy of Sciences, Prague,
June 5-7 , 1990, Prague, pp. 43-55.
Rip, Arie : 1992, 'Cognitive Approaches to Technology Policy', forthcoming in the
Proceedings of the Third NISTEP International Conference on Science and Technology
Policy Research, Oiso (Japan), 9-11 March 1992.
Rip, Arie, Schoemaker, Nelleke, and Meus, Corrie : 1983, De Universiteit als " Transf er
Agent", Leiden: LISBON, University of Leiden.
Schäfer, Wolf (ed.): 1983, Finalization in Science. The Social Orientation of Scientific
Progress, Dordrecht : Reidel.
Schot, Johan: 1991, Technology Dynamics: An Inventory of Policy Implications for
Constructive Technology Assessment, 's Gravenhage: Netherlands Organization for
Technology Assessment, NOTA Werkdocument 35.
Van den Belt, Henk and Rip, Arie: 1984, Technologie-ontwikkeling: Het Nelson-
Winter/Dosi model, Leiden : DSWO, University of Leiden.
Van den Belt, Henk, and Rip, Arie : 1987, 'The Nelson-Winter/Dosi Model and Synthetic
Dye Chemistry', in Bijker, Wiebe E. et al. (eds.) The SocialConstruction ofTechnological
Systems. New Directions in the Sociology and History of Technology; pp. 135-158.
Van den Daele, Wolfgang: 1978, "The Ambivalent Legitimacy of the Pursuit of Knowledge',
in Boeker, Egbert, and Gibbons (eds.) Science, Society and Education, pp . 23-62, VU
Bookshop, Amsterdam.
Van den Daele, Wolfgang, Krohn, Wolfgang, and Weingart , Peter (Hrsg .): 1979, Geplante
Forschung, Frankfurt a/Main : Suhrkamp.
Van Lente, Harro, and Rip, Arie: 1991, 'Membrane Technology: An Example of the
Dynamics of Strategie Scientific-Technologieal Fields', revised version of paper presented
at the International Workshop on Network Approaches in Technology Studies,
University of Groningen 30-31 May 1990, Enschede: University of Twente, Department
of Philosophy of Science and Technology .
Vergragt, Philip, Mulder, Karei, Rip, Arie, and van Lente, Harro: 1990, De matrijs van
verwachtingen ingevuld voor de polymeren Tenax en Twaron, Groningen/Enschede,
University of Groningen/University of Twente, NOTA .
Vermeij, Erwin : 1990, Contextuele verschillen in de ontwikkeling van technische
toepassingen van methaangisting, bestudeerd m.b. v. een variatie-selectie model, TWIM-
studies No . 9, Eindhoven: Technieal University Eindhoven, Research Cent re for
Engineering Sciences, Innovation and Society.
270 ARIE RIP
Vincenti, WaIter G. : 1991 , 'The Scope for Social Impact in Engineering Outcomes: A
Diagrammatic Aid to Analysis', Social Studies of Science 21, pp. 761-767.
Weingart, Peter: 1978, 'The Relation between Science and Technology - A Sociological
Explanation', in Krohn et al. (eds.) The Dynamics ofScience and Technology; Sociology
of the Seiences Yearbook, pp. 251-286, Dordrecht: Kluwer Academ ic Publishers.
Wittrock , Bj örn and Elzinga , Aant (eds.): 1985, The Unlversity Research System . The Public
Policies of the Home of Scientists, Stockholm: Almq vist and Wiksell International.
Zilsel, Edgar: 1976, Die sozialen Ursprung der neuzeitlichen Wissenschaft, Krohn , Wolfgang
(H rsg .) , Frankfurt a/Main: Suhrkamp.
Zirnan, John: 1976, The Force of Knowledge. The Scientific Dimension of Society,
Cambridge: Cambridge U.P .
Ziman, John: 1983, 'The Collectivization of Science', Proc. Royal Soc.• B 219, pp . 1-19 .
Zirnan, John: 1990, 'Research as a Career' in Cozzens, Susan E. et al. (eds.) The Research
System in Transition, pp . 345-359 .
NAME INDEX
Aaboe, A. 202 Bjerknes, V. 180, 191-201,203

Abernathy, W.J . 241,262,266 B1ackwe1l, B. 230
Abramson , A. 126-128 Blanken, I. 127
Aitken, H.G.J . 22,34,263,266 Bledstein, B.J. 226
Albers 127 Bloch , M. 259
Alders , J .C . 128 Blume, S. 204
Alter, P. 151 Boeker , E. 269
Alves, G .E. 68 Boerman, J .R . 121,130
Andel , P. van 125, 128 Böhme , G. 125, 128,259-260,263, 266
Aristotle 3-6, 13-14, 65 Bolton, W. 149
Arrnington 221 Booker, P.J . 63
Arris, R. 60 Borda, C. de 41,63
Arthur 263 Boucher, E.F. 68
Boume, J . 176
Baird , J .L. 100, 104-105 , 123, 126 Bouwman 121, 123
Barne s, B. 61,97,201,260,266 Bourdrez, J .P. 129
Barnumsty1e, P.T. 214 Bowring, J. 226
Barraclough, G . 8 Boyle 77-78,94
Basa1la 265 Braun 104, 107, 125
Baumeister, T. 60,61 ,64-65 Bridgman, P.W. 59,61,68
Becker, H .A . 57-58,61,67 Broder, A. 153
Becquey, M. 202 Bromberg, J . 201
Beer , J.J . 133, 151 Brown, C. 60
Bekker, H. 163 Bruin ing, H . 121, 126, 128-130
Belidor 182 Bruneau , L. 153
Be1l, D. 261 Brunot, A. 202
Belt, H . van den 242,260-262,269 Bucciarelli, L.L. 202
Bentham, J. 207-211,213,215-216,222, Buchanan, R.A . 202
224-227 Buchholz, K. 264, 266
Berg, S. 125, 128 Buckingham, E. 55-57, 65, 66-67
Berger, J.F. 129 Burn s, R.W. 128
Berggren, J.L. 202 Bunt, E.A . 62
Bernouille 176 Byatt , I.C.R . 134, 151
Berthillier, M. 126, 128
Bezza, B. 151, 153 Cahan, D. 137, 152
Bijker, W.E. 13, 15, 33-34, 260, 262, Ca1lon, M. 230, 241, 260-261 , 265-266,
266, 268-269 268
Billington, D. 64 Calvert, M.A . 175
Biot 89 Camerer, R. 54,66
271
272 NAME INDEX
Cambell 127 Daey Owens , C. 264,266

Cambon, V. 152 Daniels, G.H . 264, 266
Campbell Swinton, A.A. 104, 126, 128 Daugherty, R.L. 60, 66
Card, O.S . 265 David, P .A. 263,266
Cardwell, D.S.L. 64, 70, 73, 93, 97 Dawson, V.P . 66, 68
Cardot, F. 153 Degas , R. 62
Carnot 72, 74, 83-85, 88, 90, 95-97, 171, Delprat, J.P . 171,176
187,263 Deprez , M. 139
Casimir, H.B.G . 121, 124, 126-128 ,259, Derrick, R.O. 228
265-266 Descartes 62
Cauchy 181, 183 Dessauer 10
Ca yley, G . 20 Dhombres, J. 202
Chain, E. 234 Dickinson 73,93,97
Channell, D.F . 96-97 , 202 Dippel 119
Chatel ier, H . le 135-136, 151-152 Disco, C. 175-176,259,263-264,267
Chick, A.C . 64 Dittlof Tjassens, J .A . 176
Chladni, E.F . 182 Dobrowolsky, M.D . 139
Chu , E.W. 266,269 Doorn, A.G. van 128
Chubin, D.E . 266, 269 Dorsman 127, 130
Clapeyron 263 Dosi, G. 242,259,261-262,264,267,269
Clark, K.B. 241,262,264,266 Drabkin, I.E. 61
C1ausius 64,73,93,97,168-169,171 ,263 Drake, R.M. 50-51 ,64-65 ,67
Cohen, M.R . 61 Drift, A. van der 121,128
Cohen Stuart, L. 171 Druy vesteyn, M.J . 105-106, 126, 130
Cole , G.S. 266 Dublin, P.T. 67
Collier 229 DuBoff, R.B. 227
Colombo, G. 136, 146, 152 Dugac , P. 202
Con stant, E.W . 19-21 ,34,261 ,266 Duncan , W.J . 61
Cooper, C.C. 226 Dup in, C. 202
Coquand, R. 202 Durand, W.F. 61
Cornish, J.J. 61 Durbin, P .T . 14-15 ,226
Co stabel , P . 202
Couzen 228 Eckert, E.R .G . 49-51 , 64-68
Covello, V.T. 230 Edge, D. 97,260,266
Cozzens, S.E . 266, 269 Edison, T . 220,229
Crelle 82, 93-94 Einstein, A. 225
Cremer, Chr. 176 Elbaum, B. 151
Crosland, M . 201 Elias, N. 267-268
Crum, W. 187,202 Elliot, B. 267-268
Csorda, I.P . 126, 128 Elzinga , A. 265, 270
Culmann, C. 44-45, 63 Emanueli, L. 148
Cutcliffe, S.H . 226 Ernest Hili, F. 228
Czitrom , D.J . 227 Es, van der 264
Estuaries 66
Daele, W. van den 125, 128, 131, 259- Everson , G. 128
260, 266, 269 Ewen, S. 228
NAME INDEX 273
Ewing 94-95, 97 Gordon, L. 187-188,203

Grashof 55, 66
Fairbairn, W. 168, 187 Grattan-Guinness, I. 202
Farnsworth, P.T. 110-112,115,129 Graue, 1.C . 176
Farquharson, 1. 60 Gremmen, B. 265, 267
Ferranti, S.Z. de 141-143, 149, 153 Griess 240
Ferraris. G . 147 Guagnini , A. 12, 152
Fink, D.G. 129 Guillet, L. 152
Fischer, E. 261 Guyonneau, F.M . 73,97
Fischer , O. 261
Fix, H . 130 Haan, E .F . de 121-122, 128
Fleming, A . 104, 234 Haantjes 121, 130
Florey, H . 234 Haldane, R.B. 151
Ford, H . 215-225, 228-229 Haller, A. 151
Forest, L. de 104 Hampton, 1. 227
Forgue, S.V. 114,131 Harre, R. 260, 267
Foucault, M. 208-209, 215 Harte, J.H . 176
Fourcy, A . 202 Harvey, D. 230
Fourier, 1. 46,47-50,52,59-60,61,64, Hawkins, N. 218
66, 181, 183,202 Healey, P. 266
Fox, R. 83,95,97, 152 Heel, D. van 172
Francis, 1.B. 43,63-64 Heerema, H .H . 129
Francken, J.C . 121, 126, 128-130 Heidegger 10
Frankenstein, R. 219 He ijne , L. 121-122, 125, 127-129
FrankIin, P. 64 Heimholtz, H . von 52,65 , 138, 188, 233
Franzini, 1.B. 60, 66 Herivel,l . 95,97,202
Freeman, A. 64 Hertner, P . 134, 151
Freeman, C . 252, 264, 267 Hertz, H . 191
Freeman ,l .R . 64 Hili 73
Friedman, R.M. 200-201,203 Hills 93,97
Froude, W. 41,63 ,65 Hofmann 239, 261
Fox 12 Holm, W .A . 126,129
Hol st, G . 101, 106, 118-121 , 123,
Gabba, L. 151 126-127
Gauthey, E . 181-182, 201 Homburg, E . 7-8, 14-15, 93, 259, 261,
Gavroglu, K. 97 267
Gendt jr . 1.G. van 176 Hooff, W .H .P .M . van 175
Germain, S. 181-182 Hoogstraaten, P . 266
Gerstenfeld, A . 268 Hoover 229
Gibbons 269 Horowitz 229
Gillispie, e.e. 97,202 Huet, A . 172, 176
Goffman, E. 212, 227 Hughes, T.P. 6,14-15,34,141,153 ,228,
Goldstein, B.R . 202 252, 262, 264, 266-267
Goldstein, R.l . 50 ,65,68 Huning, A . 14-15
Gooday, G . 137,152 Huntley, H .E . 61,64
Goodrich, R .B. 114, 131 Hutton, C . 39, 62
274 NAME INDEX
Kranakis, E. 12-13 , 201-202, 263, 265,

lams, H. 111,126,129-130 267
Imelman, N.A . 170,175 Kranzberg , M. 226
Ince, R. 63,65, 153 Kroes, P . 12-13, 15, 58, 67-68, 96, 97,
Irvine, J. 265, 267 175,263,267
Isherwood, B.F. 42 Krohn, W. 68, 125, 128, 130, 259-260,
266-267,269-270
Jackson, J .D . 66 Krug 83, 95, 97
Jackson, P. 227 Kuhn , T.S. 34,240,267
Jack son Lears, T .J. 228 Küppers, G. 68
Jacob, E. 66 Kuyper, G. 176
Jacob, M.J . 229
Jacobi, C.G .J . 59-60 Lacaita, C.G . 152
Jakob, M. 49-50, 54, 62-66 Laeey, R. 229
Janich, P . 263, 267 Laeroix, S.F. 202
Jaspers 10 Lakatos, I. 97, 240
Jelsma, P. 266 Landes,D.S. 8, 133, 151,205-206,225
Jenkin, F. 186-187 Langhaar, H.L. 56,61, 67
Jenkins, C.F . 100, 104-105, 126, 129 Laredo, P. 265,268
Jhally 228 Laseh, C . 228
Johnson, K. 20 Latour, B. 226,241 ,244,259-261,268
Jona, E. 148 Laudan, R. 34, 266
Jones, J . 223-224 Law , J . 230, 262, 266, 268
Jongh , G.J .W. de 176 Layton J r. E.T. 12, 27, 34, 51, 61-63,
Jongkees, A. 176 65-66, 68, 69, 93, 97, 155, 175-177,
179, 201, 225, 260, 266-267
Kantrow , A.M . 266 Lazoniek . W. 151
Karman, T. von 41, 63 Leblane , X. 127, 130
Karpman, B. 227 Lee, J .R. 218
Kasson , J .F . 212,227 Leeuwin, M. 129
Kekule 261 Leiss 228
Keller, E.F . 226 Lenk , H . 14-15
Kelvin 44, 137-138, 152, 180, 185, Lente, H. van 261, 265, 269
200-203 Lentz Woodruff, L. 221
Kerker 73,81,93-97 Les1ey, E.P . 61
Kerkhoff, F . 113, 126, 129 Levine, L.W . 212,227
Kim, B. 266 Lewin 126
Kinder , W. 68 Lichtenbelt , A.D.F.W. 176
Kline, S.J. 56, 60-61, 64-65, 67, 228 Lienhard, J. 54, 66, 68
Knoblauch, O. 54 Lincoln, A. 220
Knorr-Cetina, K.D . 267-268 Lintsen, H.W. 12,93,175-176
Kodama, F. 265,267 Lohr, L. 221,229
Kohler, R.E. I, 13 Loringer 221
Kohlrausch, F. 137, 152 Lude, K. 229
König, W. 152 Lundgreen, P. 156, 175
Koyr, A. 62
NAME INDEX 275
Maartens, J.H.J. 121,130 Moser, S. 14-15

Macagno, E.O. 47,61,64-65,67 Mowery,D.C. 133, 151
MacGregor, C. 60 Mu1der, K. 261-262 ,268-269
Mach 65 Mulkay , M. 267-268
Machiavelli 244, 260 Murdock 65
Mackey, R. 14-15 Murphy, 1.G. 227
MacLeod, R.M . 151 Myers, S. 61
Madaus, G.F . 227
Maillart, R. 64 Napier , J .R. 187
Mar, B.W . 268 Navier, C.L.M .H . 51-52, 180, 182-185,
Marchand, R. 215, 228 191, 196-202,204,263
Marelli, E. 148 Nelson, R. 238,259-261 ,267-269
Mark, J . van der 109,126-127,131 Nevins, A. 228, 229
Marquis, D.G. 61 Nernst, W. 138
Martin, B.R . 265,267 Newcomen 74,76, 166
Martins, H . 267 Newell, W.T. 268
Matschoss 73, 93, 97 Newton, I. 4,38-41,44,47 ,51,62-63,
Maxwell 189-190, 198 65,86, 186, 188,225
May, 1. 115 Neyret-Brenier 145
Mayer, E. 267 Nierstrasz, 1.L. 170, 175
Mayo, B.J. 129 Nipkov, P .G. 100-107 ,118,123,126,130
Mayr, O. 13, 34, 92, 96-97 Noir, D. 65
McCallum, D.B . 230 Norri s, C. 228
McCormmach, R. 265,268 Norton Wise, M. 152
McDowell, D.M. 66 Nusselt, W. 54-55,66-67
McGinn, R. 34
Mcllwa in, K. 129 Olesko , K.M. 152
McLuhan 227 Orwell , G. 210
McMillan , F. 259, 268 Ouden, A. den 175
McPar1an , J . 227 Owens, L. 64
Menninger, K. 209, 227
Mercx, J .A. 163 Pace y, A.J . 96-97
Mertz, C. 143 Palmer, P . 230
Metivier, M. 202 Pambour, F.M.G. 70-97 , 166-169, 171,
Meulen, B. van der 259, 263, 267 176,263
Meus, C. 259, 269 Pavese, C. 153
Meyer, S. 228 Pavitt, K. 264, 268
Mitcham, C. 14-15 Pavlova, M.T. 230
Mohr, O. 44 Payen, J. 93, 95, 97
Molella, A. 13, 15 Perez 252
Menge, G . 44 Perkin 239
Moog, O. 216, 220, 228 Perrin 104
Morin 94 Perry, J . 44,64
Morris , H . 227 Philips, A.F. 129
Morse, S. 213,227 Philo of Byzantium 37,61
Morton, G.A. 126, 131 Pinch , T.J . 13, 33-34, 262, 266, 268
276 NAME INOEX
Pipijn, G. 176 Rouse 63,65

Pirelli, A. 153 Roussin 240
Plato 3-6 RusselI, B. 227
Poisson, S.O . 181, 183, 191,202 Rutherford, E. 14
Pol , B. van der 106, 129 Ryle, G . 27,34
Polanyi, M. 20-21,34
Polio, M.V . 62 Sabbah, C.A . 115
Poncelet 94 Sacchetto, H. 153
Pope, O. 228 Sackville-West, V. 35,60
Po st, R.C. 226 Sandman, P. 230
Prakke, F. 259, 266, 268 Sarlerni jn, A . 12-13, 15,58,67,93, 125-
Prandtl, L. 54-55, 66 128, 130
Prins, G. 265,268 Sarnof, O. 130
Prony 93 Savery 85
Puetz, J . 130 Saxberg, B.O . 268
Schäfer, W . 269
Querrien, A . 202 Schagen, P . 121, 126, 128-130
Schampers, P.P .M . 117, 121, 128
Rankine , W .J .M . 73, 94-95 , 97, 168, Schmidt, E . 54, 66
187,202-203 Schneider 144
Rapp, F. 10, 14-15 Schoemaker, N. 259,269
Rathenau, E. 152 Scholl, E .F . 176
Rayback, J . 227 Schot, J . 269
Rayleigh 52-53, 56, 61, 65 Schroeder van der Kolk, H .W. 176
Reech, F . 41,63 Schudson, M. 228
Regnault, V. 137, 152, 168, 171, 186 Schumpeter 242
Reich, L.S . 137, 152 Schut, T .G . 121
Reimer, G . 97 Schwarzlose, R. 227
Reingold, N . 13, 15 Scot 262
Reuleaux 263 Sedov , L.I. 61
Reuther, W . 219 Semenza, G . 147, 153
Reynolds, O . 52-55,57,65-66 Shaiken, H. 210,227
Reynolds, T.S . 63, 65 Sheeler, C. 229
Rice, E . 137 Shepherd, O.G . 66
Rijssel , T .W . van 121,130 Sheridan, A . 227
Rinia, H . 118-119, 121, 123, 126-127, Shinn, T . 202
130 Shoenberg, I. 110
Rip, A . 12, 259-269 Siemens, G. 152
Rivera, O. 229 Sigurdson, J. 267
Rivers 66 Silverberg, G . 267
Rond d' Alembert, J . le 39 Simon, H .A . 21-22,28 , 34
Rorty, R. 228 Sims 221
Rosaldo, R. 226 Singer, C . 73,93
Rose, A . 111,126,129,130 Skinner, B.F. 227
Rosenberg, N. 13, 15, 137, 152,201 Smeaton, J. 41,63 ,88
Rossini 212-213 Smith, A . 206
NAME INDEX 277
Smith, C. 137, 152, 186-188,200-203 Velde, T.S. te 125

Smith, J .M. 130 Verbong, G. 93
Soete, L. 263, 267 Verdam, G.J . 171, 176
Solla Price, D. de 231,259-260,268 Vergragt, P . 261-262, 268-269
Southern, J . 44 Vermeij, E. 264, 269
Starnberger 125 Verwey, E. 121
Staudenmaier, J.M. 2, 8-12, 13-15, 69, Vincenti, W.G . 9, 14,34,61-62,69,93,
93,97 96-97 ,201,225 ,263-264,270
Steen, A. van der 176 Vitruvius 38
Steenaard, R. 175 VIissingen, P . van 172
Stern, R.H . 130 Vries, M. de 12,93, 126
Steward , J . 130
Stieltjes 130 Wagner ,D .J . 176
Stokes, G.G. 51-53,65 Walker Lurito, P . 228
Ströker,E.II,14 Warneeke, H .J . 268
Strutt, J .W. 52,65 Watt, J . 44,64, 73, 88, 90, 97, 166
Webster, N. 220
Teer, K. 130 Weimer, P .K. 114,131
Tempel Bell, E. 68 Weingart, P. 68,260,266-267,269-270
Tesla , N. 139 Weingartner, P . 131
Thackray, A. 3,226 Weisbach, J . 44,63, 172, 176
Thiele, R. 130 Wengenroth, U. 139, 152-153
Thornas, T. 213 Werner , W. 113, 126, 129
Thompson, R.L. 227 Westermann, F. 228
Thom son , J . 187-188,203 Whipp en, W.G. 60
Thom son , W. 104, 180; 185-191, Whitley, R. 267
196-204 Whitney, W.R . 152,204
Tiltman, R.F. 126, 130 Wightman Fox, R. 228
Timoshenko, S.P . 63,202 WilIs, C.H . 218
Torricelle 43 Winter, S. 238,259-261, 268-269
Toulm in, S. 226, 230 Wise, G. 152
Touraine 261 Wise, M.N . 137, 187-188,200-204
Tower, B. 53,66 Wittrock , B. 265, 270
Towler, G.O . 99, 125-126, 130 Woolgar , S. 259, 268
Tozer, L. 229
Treue , W. 152 Zeuner, G.A . 97, 168
Trevor, J . 266 Ziegler 234
Zilsel, E. 259,270
Upward, G.C . 228-229 Ziman , J . 259, 265-266, 270
Zuboff, S. 210,226-227,230
Varley, C .F . 186 Zworykin, V.K. 106,109-110,126,131
Vaschy, A. 55-57 ,66-67
Boston Studies in the Philosophy of Science
121. P. Nicolacopoulos (ed.): Greek Studies in the Philosophy anti History 0/
Science.1990 ISBN 0-7923-0717-8
122. R. Cooke and D. Costantini (eds.): Statistics in Science . The Foundations of
Statistical Methods in Biology, Physics and Economics. 1990
ISBN 0-7923-0797-6
123. P. Duhem: The Origins 0/ Statics. Translated from French by G.F. Leneaux,
V.N. Vagliente and G.H. Wagner. With an Introduction by S.L. Jaki, 1991
ISBN 0-7923-0898-0
124. H. Kamerlingh Onnes: Through Measurement to Knowledge . The Selected
Papers, 1853-1926. Edited and with an Introduction by K. Gavroglu and Y.
Goudaroulis. 1991 ISBN 0-7923-0825-5
125. M. Capek: The New Aspects 0/ Time: Its Continuity anti Novelties. Selected
Papers in the Philosophy of Science. 1991 ISBN 0-7923-0911-1
126. S. Unguru (ed.): Physics, Cosmology anti Astronomy, 1300-/700. Tension and
Accommodation. 1991 ISBN 0-7923-1022-5
127. Z. Bechler: Newton's Physics on the Conceptual Structure 0/ the Scientific
Revolution. 1991 ISBN 0-7923-1054-3
128. E. Meyerson: Explanation in the Seiences. Translated from French by M-A
Siple and DA Siple. 1991 ISBN 0-7923-1129-9
129. AI. Tauber (ed.): Organism and the Ortgins 0/ Self. 1991
ISBN 0-7923-1185-X
130. F.J. Varela and J-P. Dupuy (eds.): Understanding Origins. Contemporary
Views on the Origin of Life, Mind and Society. 1992 ISBN 0-7923-1251-1
131. G.L. Pandit: Methodological Variance. Essays in Epistemological Ontology
and the Methodology of Science. 1991 ISBN 0-7923-1263-5
132. G. Munevar (ed.): Beyond Reason. Essays on the Philosophy of Paul
Feyerabend.1991 ISBN 0-7923-1272-4
133. T.E. Uebel (00.): Rediscovering the Forgotten Vienna Circle. Austrian Studies
on Otto Neurath and the Vienna Circle. Partly translated from German. 1991
ISBN 0-7923-1276-7
134. W.R. Woodward and R.S. Cohen (eds.): World Views and Scientific Discipline
Formation . Science Studies in the [former] German Democratic Republic.
Partly translated from German by W.R. Woodward. 1991
ISBN 0-7923-1286-4
135. P. Zambel1i: The Speculum Astronomiae anti lts Enigma . Astrology, Theology
and Science in Albertus Magnus and His Contemporaries. 1992
ISBN 0-7923-1380-1
136. P. Petitjean, C. Jami and A.M. Moulin (eds.): Science anti Empires. Historical
Studies about Scientific Development and European Expansion.
ISBN 0-7923-1518-9
137. W.A Wallace: Galileo' s Logic 0/ Discovery and Proof. The Background,
Content, and Use of His Appropriated Treatises on Aristotle 's Posterior
Analytics. 1992 ISBN 0-7923-1577-4
Boston Studies in the Philosophy of Science
138. W.A. Wallace: Ga/i/eo's Logical Treatises. A Translation, with Notes and
Commentary, of His Appropriated Latin Questions on Aristotle's Posterior
Analytics.1992 ISBN 0-7923-1578-2
Set (137 + 138) ISBN 0-7923-1579-0
139. M.J. Nye, J.L. Richards and R.H. Stuewer (eds.): The Invention 0/ Physical
Science. Intersections of Mathematics, Theology and Natural Philosophy since
the Seventeenth Century. Essays in Honor of Erwin N. Hiebert. 1992
ISBN 0-7923-1753-X
140. G. Corsi, M.L. dalla Chiara and G.C. Ghirardi (eds.): Bridging the Gap:
Philosophy, Mathematics and Physics. Lectures on the Foundations of Science.
1992 ISBN 0-7923-1761-0
141. C.-H. Lin and D. Fu (eds.): Philosophy and Conceptual History 0/ Science in
Taiwan.1992 ISBN 0-7923-1766-1
142. S. Sarkar (ed.): The Founders 0/ Evolutionary Genetics. A Centenary Reap-
praisal. 1992 ISBN 0-7923-1777-7
143. J. Blackmore (ed.): Ernst Mach - A Deeper Look. Documents and New
Perspectives. 1992 ISBN 0-7923-1853-6
144. P. Kroes and M. Bakker (eds.): Technological Development and Science in the
Industrial Age. New Perspectives on the Science-Technology Relationship.
1992 ISBN 0-7923-1898-6
145. S. Amsterdamski: Between History and Method. Disputes about the Rationality
of Science. 1992 ISBN 0-7923-1941-9
146. E. Ullmann-Margalit (ed.): The Scientific Enterprise. The Bar-Hillel ColIo-
quium: Studies in History, Philosophy, and Sociology of Science, Volume 4.
1992 ISBN 0-7923-1992-3
147 L. Embree (ed.): Metaarchaeology, Reflections by Archaeologists and Philos-
ophers. 1992 ISBN 0-7923-2023-9
Also 0/ interest:
R.S. Cohen and M.W. Wartofsky (eds.): A Portrait 0/ Twenty-Five Years Boston
Colloquia for the Phi/osophy ofScience , 1960-1985. 1985 ISBN Pb 90-277-1971-3
Previous volumes are still avai/able .
KLUWER ACADEMIC PUBLISHERS - DORDRECHT / BOSTON / LONDON

Technological Development and Science in The Industrial Age

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

Technological Development and Science in The Industrial Age

Uploaded by

Copyright:

Available Formats

TECHNOLOGICAL DEVELOPMENT AND SCIENCE

IN THE INDUSTRIAL AGE

ROBERT S. COHEN, Boston University

Editorial Advisory Board

THOMAS F. GUCK, Boston University

Springer-Science+Business Media, B.V.

Teehnologleal develop.ent and selenee In the Industrlal Age : new

ISBN 978-90-481-4186-9 ISBN 978-94-015-8010-6 (eBook)

Printed on acid-free paper

All Rights Reserved

PETER KROES and MARTIJN BAKKER I Introduction:

WALTER G. VINCENTI I Engineering Knowledge, Type of

EDWIN T. LAYTON Jr. I Escape from the Jail of Shape;

PETER KROES I On the Role of Design in Engineering

ANDRIES SARLEMIJN and MARC DE VRIES I

JOHN M. STAUDENMAIER, S.J . I Science and Technology:

ARIE RIP I Science and Technology as Dancing Partners 231

NAME INDEX 271

This book is the offspring of a conference on the relationship between

ing science from technology received especially close attention; it was

somehow not reliable indicators for science and technology. Nowadays, it

Within the Aristotelian philosophy, science in the strict sense (that is

phenomena, and it seems for all other sciences. It is interesting to point

It is rather paradoxical to observe that, precisely in the era in which the

technology relationship than with the nature of technological knowl-

characteristic features of technological knowledge (as opposed to

notions of technology and science play an important role, whereas these

the hope of demystifying the distinction and of finding new, fruitful

interpretation schemes and models of the interaction between science and

of Technology and Culture; see Reingold and Molella (1976).

10 Staudenmaier (1985), p. 85.

11 See Kranakis's contribution to this volume .

in Sarlemijn and Kroes (1990), pp . 3-21.

the 16th and 17th centuries.

21 Hughe s (1989), p. 123.

of physics pure and simple" (quoted in Hughes (1989), p. 123).

24 See Homburg (1986), p. 373.

" Homburg (1986), p . 368; the translation is ours ,

27 Staudenmaier (1985), p. 120.

28 Staudenmaier (1985), p. 103.

genuine methodological innovation on the topic" .

31 For a survey of the philosophy of technology, see C. Mitcharn , 'Philosophy of

Technology' in Durbin (1980), pp . 282-363 .

Rapp (1974); Huning (1974); Rapp (1981).

3 4 Rapp (1974), p. xii.

science develops and progresses.

in Durbin and Rapp (1983), p , 333.

ENGINEERING KNOWLEDGE, TYPE OF DESIGN,

This article comes at a point in my work that is both advantageous and

To the extent that scholars concern themselves with technological

Philosophical analysis of technological knowledge shows correspon-

to be defined for present purposes as the practice of organizing the design,

NORMAL AND RADICAL DESIGN

When we examine knowledge in the context of everyday design, ideas

more stringent conditions''' . 4 The designers whose needs I was studying

USE AND GENERATION OF KNOWLEDGE

- - ~1e-;;lIi1c k-;;O';Ied~- - 7/ 7/,7h'/ /7-0;- Engl"';rln; k;'wle,j;.- - -

Fig. 1. Knowledge and its generat ing activit ies,

my book I attempt in effect to unpack the contents of the solid-line box

Another concept, basic to design though not reflected in the diagram, is

Another point is essential: Whether design is radicalor normal is

commercial and military customers to obtain airplanes with tlying

Layton's words, develop "their own bodies of more or less systematic