'The Linear Model' Did Not Exist: Reflections on the History and Historiography of Science and Research in Industry in the Twentieth Century

Full text

‘The linear model’ did not exist: Reflections on the history and historiography of science
and research in industry in the twentieth centuryi
David Edgerton
in Karl Grandin and Nina Wormbs (eds), The Science–Industry Nexus: History,
Policy, Implications. (New York: Watson, 2004)
‘The linear model’ has become a term of art in studies of science policy and innovation, and
in some historical studies of science and technology.ii It is, like ‘technological determinism’
and ‘Whig’ history of science and technology, an invention of academic commentators.iii Like
these, but unlike ‘scientific revolution’ or ‘big science,’ ‘linear model’ was not meant to be an
analytically useful concept: it is there to be condemned as simplistic and inaccurate. It is a foil
for the more elaborated academic account, in short, a classic straw man. But it is more than a
straw man: although it is of recent invention, some students of science and technology have
given the model historical agency. They have come to believe that it existed in the minds of
academic analysts and key policymakers of the past, and that it had a powerful influence on
policy and practice. Worse still, the idea of ‘the linear model’ often locks even critics into a
concern with ‘basic’ science, even in the study of ‘innovation’: proponents (such as they are)
and critics, share a model of science in which science is academic research. In this model
studies of academic research are privileged, as is innovation in such studies.
I will argue that using and criticizing the term ‘linear model’ avoids critical engagement with
the much richer models of innovation developed by academic specialists in innovation, as
well as many crucial historical actors. Accounts of innovation in the 20th century, and indeed
science in industry in the 20th century, more usefully start from a conceptual frame quite
different from either the ‘linear model’ or the usual criticisms of the model. In particular the
history and historiography of non-academic research is a key resource. For example, the
history of industrial research and of science in industry, and new accounts of military research
and development—both significantly often treated as part of the history of ‘technology’—
provides a rich alternative reading of the history of twentieth century science, including the
development of academic science, ‘big science,’ interdisciplinary research, and more
obviously the ‘industrialization of research,’ which historians of ‘science’ should pay attention
to.iv We need to be careful, however, because the academic research model has affected even
our understanding of science in industry and the military. Industrial and military research and
1

development was much more than central corporate or government research laboratories. I
argue that we should go further still, and note the systematic conflation between ‘science’ and
‘research:’ most research is not academic, and most science is not research. Finally, I will
argue that the case of the ‘linear model’ allows us to reflect on the general tendency to attack
straw men in academic studies of science and technology, and on the lack of cumulation in the
historiography of science and technology.
What is ‘The Linear Model’?
‘The linear model’ is clearly a term of art, but one that is rarely if ever closely defined. The
term tends to be used in the sense of a model of innovation, rather than say, science, but most
commonly it is a model of the interaction of science and society, and science and economic
performance specifically. The model is often presented diagrammatically, as in Figure 1, and
comes in many variants. The lack of clarity, the lack of consensus, or indeed debate over the
details of the model is itself indicative that we are not dealing with a worked-out model which
anyone ever believed in. Yet we can usefully distinguish some common themes. The ‘linear
model’ incorporates three elements—the nature of the sources of innovation, of the
innovative process, and the effect of innovation. ‘The linear model’ is usually taken to be
something like the following: ‘basic’ or ‘fundamental,’ ‘pure’ or ‘undirected,’ scientific
research is the main source of technical innovation; the process of innovation is a sequential
one, by which discoveries arising in such research are developed in a sequence through
applied research, development and so on, to production. Overall, the innovation produced is
the main source of economic growth.v This reading is very close to what Donald Stokes takes
the ‘linear model’ to be, in his recent study of science policy, as I discuss further below.vi
Although is it is often described as a ‘linear model of innovation’, even when described in this
way, it includes a model of effects of innovation: For example, Harvey Brooks describes the
“linear-sequential model of technological innovation in which radical innovations are
triggered by new scientific discoveries and become foci for the growth of new industries and,
thereby, sources of economic growth and employment.”vii Yet stated as clearly as above the
‘linear model’ is very hard to find anywhere, except in some descriptions of what it is
supposed to have been. The most brazen propagandist for scientific research would wish to
avoid formulations which so explicitly beg so many questions: ‘the linear model’ not only did
not exist, but it could not exist as an elaborated model.
2

[FIGURE?]
That the ‘linear model’ has been significant is not itself a straw man. By the 1990s ‘the linear
model’ and ‘the linear model of innovation’ were terms in very widespread use in the
academic and official literature (as a check with a search engine, or leafing through the pages
of specialist journals like Research Policy will verify). It is always, however, something that
was surpassed, criticized, to be moved beyond.viii That there was widespread discussion of the
‘linear model’ was often commented on in academic literature in the 1990s. In 1993 I claimed
that “there is much condemnation of the so-called ‘linear model’ of innovation: the argument
that science, leads to technology, and thence to economic growth.”ix By 1995 one paper in
Research Policy was claiming, “The linear model of innovation is the key reference point for
understanding the relationship between science, technology and economic development.”x
The doyen of innovation studies, Chris Freeman, writing in 1996, noted that “at one time it
was almost impossible to read a book or an article on technology policy or technological
forecasting that did not begin or end” with a polemic, against the “so-called ‘linear model of
innovation.’”xi Freeman himself suggested that “The linear model cannot […] be dismissed
simply as a convenient straw man erected for the convenience of those expounding alternative
ideas.”xii Ernest Braun, another veteran science policy academic, and around the same time,
complained about the tendency to attack the crudest linear models, which in effect no one had
ever believed in.xiii
‘The linear model’ is a term of art without a history. As far as I know no-one has enquired into
the origins of the term ‘linear model,’ though the idea that is a recent creation as implied in
one useful account of the concept in the context of ‘science parks:’ Doreen Massey and
colleagues see the ‘linear model’ as used in Britain as something recent, developed out of a
particular historical account of science and industry in British history.xiv If the term has no
history, the underlying concept, has a very elusive one. Many accounts imply and some state
(as I show further below), that the model was central to Vannevar Bush’s Science: The endless
frontier, and that this was the key to the influence of the model. A recent example is found in
an essay review in Isis where it is noted that “historians have grown skeptical of the
interpretative reliability of the so-called ‘linear model’ of ’science-push’ innovation, which, as
popularized by Vannevar Bush and others, became an axiom of faith for many who drove and
defended science and technology policy for over fifty years.”xv
3

Origins
The earliest use of the term ‘linear model’ in the context of innovation appears to be by
William J. Price of the US Air Force Office of Scientific Research, and L.W. Bass of Arthur
D. Little, in 1969. They argued, in Science, that
Innovation is often viewed as an orderly process, starting with the discovery of new
knowledge, moving through various stages of development, and eventually emerging
in final, viable form. According to this “linear” model, innovation seems to be a
rational process, essentially similar to the other, more systematic functions of an
organization. The assumption is that it can be analyzed into component parts and
controlled rationally—that is to say, planned, programmed, managed much as other,
more routine activities are.
By contrast they argued that studies of innovation showed that the “‘linear’ model” was not
typical, that the innovation process was “irrational” and could not “be programmed in
advance.”xvi If this is indeed the origin of the term, then from the first it is a term of criticism
of a particular set of views of innovation. Interestingly, the view of innovation criticized is
richer than the later ‘linear model,’ and so indeed was their criticism. They observed that in
the Second World War, researchers brought many innovations forward. These researchers,
were acting as ‘technologists,’ and not, they argued, as was often suggested, as ‘basic
researchers.’ They also criticized the standard postwar view that innovation was a rational
ordered plannable process. This view, it should be noted, was most radically and interestingly
outlined by Schumpeter, and much criticized by later neo-Austrian economists with interests
in science, like John Jewkes.xvii
The second use of the term that I know is also specifically concerned with innovation, and is
also richer than later versions, though in another way. The once well-known British collection
of case studies of innovation, Wealth from Knowledge, published in 1972, discusses and
criticizes “linear models of innovation.” It takes then to be of two very different types neither
of which corresponds to the idea of ‘the linear model’ in use today. The authors distinguished
between the “discovery push” and the “need pull” “linear models,” and then broke them down
further into four in total: 1) the “science discovers, technology applies” model, 2) the
“technological discovery” model 3) the “customer need” model and 4) the “management by
objectives model.” They go on to claim that few of their cases fit any of these models; they
4

criticize them all.xviii The specific idea that basic or fundamental research was the main source
of innovation (model 1) they saw as “widely held,” and often regarded as “self-evident” but
they noted how rarely substantiated claims were made for it.xix Edwin Layton, citing Wealth
from Knowledge, noted in the mid-1970s that there were a “variety of models” of technical
development, but that “all of these models postulate linear-sequential models of the
innovative process,” in which the linear sequence of cause and effect, followed from the first
event. He agreed strongly with Wealth from Knowledge that actual innovation did not proceed
in this linear-sequential way.xx By 1985, according to one study, there were two “traditional”
models, the demand-pull and supply push.xxi
However, the use of the terms ‘the linear model’ or many ‘linear models’ appears to have been
very rare indeed before the mid-1980s. It is particularly significant that it is not used in most
of the important work on innovation of the time. Nathan Rosenberg does not seem to use the
terms at all in his seminal collection of papers from the 1960s and 1970s Perspectives on
Technology.xxii Although the nature of innovation, and its sources, are central themes
Rosenberg discussed the work of Schmookler and Schumpeter, Marx and Engels, and many
other economists and economic historians, none of whom were anywhere near endorsing ‘the
linear model’ not least because they adopted much richer positions. Chris Freeman’s standard
text, The economics of industrial innovation (second edition, 1982), doesn’t mention ‘linear
models’ either. This is especially relevant because he was concerned not just with industrial
innovation, but also government policy. Freeman indeed divides the post-war years into two
phases of government “science and technology policy”: the first had a strong “supply side”
emphasis on “building up strong R&D capability.” He very clearly has in mind not pure
science, or academic research, but mainly large-scale R&D projects military and civil in
principally in the atomic and aeronautical fields. Freeman’s second phase dates from the late
1960s had a stronger ‘demand’ orientation, and was influenced, as he sees it, by economic,
political and environmental critiques of the supply side approach.xxiii
Nor is the ‘linear model’ there in more detailed work on these topics in the pioneering days of
the journals Science Studies, or Research Policy, which started publication in the early
1970s.xxiv It is missing from the seminal 1977 paper by Nelson and Winter “In search of a
useful theory of innovation,”xxv and the well-known empirical study on the sources of
invention by Vivien Walsh.xxvi Indeed to this day much good work in historical and other
studies of innovation does not feel the need to invoke ‘the linear model.’ For example,
5

Mowery and Rosenberg’s book of 1989 does not appear to mention ‘linear models’,xxvii nor do
some more recent good textbooks on technology and society, by for example, Rudi Volti and
Kurt Jacobsen.xxviii Historical studies of US and British science policy do not generally use the
term ‘linear model,’ xxix and key reviews of the economics of science and innovation don’t
mention it either.xxx
However, since the 1980s the term ‘the linear model [of innovation]’ has become a standard
term in much of the literature, extending to historical work, to describe what is taken to be a
standard or traditional position. Use of the term diffused very quickly, together with the
assumption that it was dominant even in the academic literature on science and technology.
As already noted, it always referred to a position that had been criticized and/or should be
criticized as a myth, but was itself allegedly prevalent and influential. The first academic use
of the term in the singular, common form of ‘linear model’ dates from 1983. It is due to Stuart
MacDonald, and is as follows: “Although a staunch defense of the linear model of innovation
[…] would be rare, it is very convenient when dealing with such an uncertain process as
technological change to assume that everything hangs on research.”xxxi A paper by S.J. Kline
on industrial innovation (from 1985), credits the Price and Bass paper of 1969 with the term
‘linear model,’ referring specifically to the idea of innovation an “orderly process starting
with the discovery of new knowledge.” Kline recognizes that Price and Bass saw this as an
inadequate model, but he sees the model as central: “the linear model continues to underlie
the thinking in many current speeches and much writing,” seeing it as implicit in push-pull
models, and in the very term ‘R&D.’ He goes on to claim, “no other model has been
available.” In this case, as in others, it is unclear what exactly is being referred to since no
examples are cited, nor is the prevalence of the argument examined, nor is it clear whether the
market-pull linear model is taken as a linear model or not.xxxii But one 1988 textbook already
refers to the “notion that technical innovations result from the application of new scientific
insights and ideas. Such a notion is often referred to as the linear model of innovation.”xxxiii
The term ‘linear model’ quickly came to be used in papers concerned with new theoretical
approaches to the history and sociology of science and technology. Trevor Pinch and Wiebe
Bijker, in their well-known joint manifesto for SCOT (social construction of technology)
found it useful to attack the “widespread use of simple linear models to describe the process
of innovation.” Although they refer to models in the plural this is because “The number of
developmental steps in these models seems to be rather arbitrary […]” and varies. They
6

clearly have academic studies in mind since they suggest that they have contributed a great
deal to understanding of “conditions of economic success in technological innovation;” they
cite no examples, and claim that such models are not of interest to them because they ignore
technical content! They also claim that their “multi directional” model is preferable to the
‘linear model’, which they see as important in “many” innovation studies and “much” history
of technology. xxxiv In this latter case the ‘linear model’ does not seem to be focused on basic
research. Pinch and Bijker were not alone in making use of ‘the linear model’ as something
for new approaches to compare themselves with. Arie Rip noted in a paper given at a
conference of historians of science and technology specifically discussing science-technology
relations:
It is convenient to start with a brief discussion of the so-called “linear model” of the
relation between science and technology. That technological innovation derives from
scientific discovery, as it were in a linear sequence, is a myth, but a prevalent 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 magic wand. The linear model has some truth
in it, but it hides more than that it helps our understanding.xxxv
Did the ‘Linear Model’ exist by Other Names?
If, as seems clear, no academic study of innovation, has ever proposed or defended a ‘linear
model of innovation’ it does seem rather odd that historians of sociologists of science and
technologists of the 1980s came to believe that this had been so. Perhaps this was because a
‘linear model’ had been implicit in such literature. And yet, as we have see a generation of
researchers working in the 1960s and 1970s was clearly either indifferent to or hostile to any
implicit ‘linear model.’ But even the earlier academic literature, is not dominated by an
implicit ‘the linear model.’ Take the British economists Charles Carter and Bruce Williams
who did a great deal of empirical work on innovation in industry in the 1950s.xxxvi They
warned of the likelihood of “overestimating the value of research in industry.” In some firms
it was
too academic—too little oriented towards commercial needs. The misconception
underlying the latter waste of scarce scientific resources is that research is naturally a
7

left-to-right process—that is fundamental research produces something which is
communicated to the industrial scientist, who performs some applied research and
communicates the results to someone else who takes matters a step further. We have
not found any cases of successful industrial research where this movement was not
accompanied by movement in the opposite direction.xxxvii
Carter and Williams found in their studies that the majority of projects started outside R&D
laboratories. If there was a ‘linear model’ it was a bi-directional one, which surely doesn’t
count.xxxviii Clearer still is the case of The Sources of Invention by John Jewkes and others of
1958: it has powerful criticisms of many different assumptions that were made about many
aspects of invention. They argued that the relations between science and technology were
simply not known, that it was not obvious that the scientifically most wealthy country would
be the richest, and warned against investing in ‘pure science’ in the expectation of a pay-
off.xxxix Sources of Invention was particularly notable as a neo-liberal attack on the
bureaucratization of innovation, and the assumption that it was predictable, an important
element in one early version of ‘the linear model,’ as we have seen. More generally, in the
1960s at least, economists pointed to the clear lack of positive correlation between
expenditures on research and development and economic growth—a clear general argument
against the linear model at the macro-level.xl At the same time there was a great emphasis on
‘demand-pull’ models of innovation. It is thus difficult indeed to argue that economists put
forward, or believed in, ‘the linear model of innovation.’ They certainly distinguished
between ‘science,’ ‘invention,’ ‘innovation,’ ‘diffusion’ and so on, but that doesn’t amount to
a model of innovation. The closest I could find to an endorsement by an economist, and then
only for heuristic purposes, of something that looks like a linear model, is in a textbook by
Rosegger of the 1980s, which refers to ‘stage models.’xli On another dimension it is clear that
historians of technology rejected the model of science-technology relations implicit in the
linear model, even in the 1940s and 1950s.xlii
My claim is then, that the ‘linear model’ did not exist in even the earliest generations of
academic work on innovation. Did then ‘the linear model’ exist elsewhere, for example in the
in the writings of scientists and engineers? The academic students of innovation, going back
to the 1950s, are all criticizing something which looks a bit like the linear model. Certainly
something like it can be found, not least if we see it as a general argument in favor of the
utility of ‘theory’ as well as ‘practice,’ of ‘pure’ as well as ‘applied science’ and for the
8

importance of ‘fundamental’ and ‘basic’ research. There is little doubt, for example, that
academic research scientists have long made, and continue to make, exaggerated claims for
the significance of their work for technological and economic development, and that agencies
which came to fund them did the same. These arguments have often had the aim of securing
state support. We could, if we wished, label these obviously self-interested claims the ‘linear
model’, but to call the propaganda of academics ‘the linear model’ is to flatter the claims, and
to avoid stating the obvious: that these are generally claims by academic researchers for the
power of academic research. To call it a ‘linear model’, also runs the risk, especially in the
context of the current use of the term, of smuggling in the assumption that what academic
research scientists said about innovation was the most influential discourse of innovation
around. In other words, to believe that ideas about innovation were created by academic
research scientists, and diffused out to engineers, to government, to industry and to the public;
to believe in a linear model not of innovation, but of ideas about innovation. That second
‘linear model’ is already implicit in the much literature in the history of science, and even
technology.
What models of innovation, public and private, where used by government officials, industrial
researchers, academics, is an open historical question. It is, however, worth presenting some
tentative arguments, if only to stimulate research. Firstly, non-academic scientists and
engineers have often been resistant or indifferent to the ‘linear model’. For the US case,
George Wise was clear that the ‘assembly line model’ was the product of the ‘science-policy
elite’ and not more widespread than that.xliii In a 1946 British conference concerned with
industrial research, the closest thing to a linear model is the claim that while day to day
advances came from industrial research, “really spectacular advances” and the “creation of
new industries” came from “fundamental research,” but this was from the government official
who was charged with funding such research.xliv The papers of a conference of US industrial
research managers of 1954, reveals no explicit or implicit linear model in its deliberations.xlv
In 1960 the research heads of important US companies were presented with a question: why
was there such a lag between scientific discovery and industrial and military application in the
USA and the free world; there was a concern that the Soviet Union was doing better. The
questioners wanted ‘ways to shorten the “pipeline” between original scientific discovery and
engineering application’.xlvi The question assumed the ‘linear model’ but the answers from
directors of research, chief engineers and so on, are notable for questioning the question. The
9

lag was shortening, said some, it was inevitable said, others, they assumed the question
referred to many different processes, and so and so forth. It clear that the managers were not
thinking is anything like as simplistic a way as the question implied. The ‘linear model’ was
not the common understanding in industry.
Second, while academic research scientists’ arguments have generally been crude and
repetitious they also often showed a richer and broader understanding of the role of
academic research than the crude linear model (involving training of scientists and engineers,
for instance). Thus Vannevar Bush’s famous 1945 report Science: The endless frontier is
much more subtle and interesting. This is a key case because in academic works which refer
to ‘the linear model’ it is typically the only cited work which predates papers critical of the
model from the 1980s. Many commentators are explicit in their attribution of the model to
Bush, and/or see Bush as its most influential exponent. Chris Freeman has claimed that
Science: the endless frontier “did indeed outline a linear model of science, technology and
innovation, and […] was certainly influential among policy-makers.”xlvii More recently,
Donald Stokes notes a “postwar paradigm” exemplified by Science: The endless frontier.xlviii
He sees the paradigm for “science policy” as resting on the view that “basic science” should
be unconstrained, and that this will lead to technological innovation, which became what he
calls “the familiar ‘linear model’” when extended through to production.xlix Stokes argues that
Bush endorsed a strong form of the linear model—”basic advances are the principal source of
technological innovation [original emphasis],” noting too the use of the term “technological
sequence” by the National Science Foundation in the early 1950s.l Elsewhere Stokes notes,
“Nothing in Bush’s report suggests that he endorsed the linear model as his own,” though he
did assert that scientific discoveries are a source of technological progress.li We need to clarify
what Vannevar Bush was arguing, and not arguing, in his supposedly foundational text.lii
Another Look at Science: The Endless Frontier
Vannevar Bush was the wartime Director of the US Office of Scientific Research and
Development, one of the many US agencies concerned with warlike R&D. In FY 1945 the
OSRD spent around $100m, the US Army and Navy $700m between them and the Manhattan
Project $800m.liii His Science: The endless frontier was a proposal for very particular policies
concerned with a small part of postwar research. The report is not concerned with innovation
nor with how this should be organized, thus it could not even implicitly set out a linear model
of innovation. The term ‘linear model’ is nowhere used. Bush was arguing for the public
10

support of basicliv science in universities at a time when he thought the growth in such
research had failed to keep up with the huge rises in government and industrial research both
of which were overwhelmingly and necessarily ‘applied.’lv He did not (as many later analysts
have done when thinking in the ‘linear model’ terms they supposedly abhor) conflate policy
for academic research, with policy for innovation. Thus “Expenditures for scientific research
by industry and Government—almost entirely applied research—have more than doubled
between 1930 and 1940. Whereas in 1930 they were six times as large as the research
expenditures of the colleges, universities, and research institutes, by 1940 they were nearly
ten times as large.” And he went on to note, “expenditures for scientific research in the
colleges and universities increased by one-half during this period, those for the endowed
research institutes have slowly declined.” The war made things worse:
We have been living on our fat. For more than 5 years many of our scientists have
been fighting the war in the laboratories, in the factories and shops, and at the front
[…] they have been diverted to a greater extent than is generally appreciated from the
search for answers to the fundamental problems—from the search on which human
welfare and progress depends.
The key point was that “If the colleges, universities, and research institutes are to meet the
rapidly increasing demands of industry and Government for new scientific knowledge, their
basic research should be strengthened by use of public funds.”
It is easy to find in it claims which we might loosely take to be ‘the linear model,’ for
example, “to secure a high level of employment, to maintain a position of world leadership—
the flow of new scientific knowledge must be both continuous and substantial.” Or, “There
must be a stream of new scientific knowledge to turn the wheels of private and public
enterprise.” Or, “Today, it is truer than ever that basic research is the pacemaker of
technological progress.” Or, that the universities, and other centers of basic research, were
“the wellsprings of knowledge and understanding. As long as they are vigorous and healthy
and their scientists are free to pursue the truth wherever it may lead, there will be a flow of
new scientific knowledge to those who can apply it to practical problems in Government, in
industry, or elsewhere.” Crucially however, he never claims basic research as the main source
of invention, or innovation, and indeed he sees basic research as leading not to new products
or processed, but repeatedly to ‘knowledge’ and to ‘understanding’ (terms which I have
11

italicized in the quotations). Thus basic research “results in general knowledge and an
understanding of nature and its laws. This general knowledge provides the means of
answering a large number of important practical problems, though it may not give a complete
specific answer to any one of them.”lvi Universities were “uniquely qualified by tradition and
by their special characteristics to carry on basic research. They are charged with the
responsibility of conserving the knowledge accumulated by the past, imparting that
knowledge to students, and contributing new knowledge of all kinds.” For the US could “no
longer count on ravaged Europe as a source of fundamental knowledge. In the past we have
devoted much of our best efforts to the application of such knowledge, which has been
discovered abroad. In the future we must pay increased attention to discovering this
knowledge for ourselves particularly since the scientific applications of the future will be
more than ever dependent upon such basic knowledge.” Indeed his own model is not so much
a linear chronological one, though that element is there, but a spatial one, in two senses. First,
different kinds of scientific activity take place in different spaces, and secondly, the extension
of scientific knowledge creates a new enlarged arena for the actions of others. That is the
significance of the term ‘endless frontier’ for, as Bush noted, it was an established policy of
the US government “that new frontiers shall be made accessible for development by all
American citizens.” As Arie Rip points out, Bush had a “reservoir model” of the role of basic
science, and indeed was somewhat avant la lettre here.lvii However, the idea that what basic
academic research produced was knowledge was not unique to Bush. For example a 1931
study of US industrial research, which like Bush highlighted the relative numerical strength of
industrial to academic research, claiming a ten-fold advantage, saw the industrial applied
researchers looking to the “pure scientists for fundamental information [emphasis added]”.lviii
The same study saw “industrial research” as the “managerial means for the systematic
application to technology of the fundamental knowledge gained by pure science [emphasis
added]”.lix
In Britain there was no equivalent to the Bush report, not least because Bush’s central claim
was unnecessary: the state had long funded academic research by direct support of
universities and of university research via the so-called ‘research councils’, including the
Department of Scientific and Industrial Research (DSIR). But while there was indeed strong
support for such research after the Second World War, and it did indeed increase, it is hard to
see a ‘linear model’ in operation even for ‘basic’ research. Sir Henry Tizard, the chief
government scientific adviser for both the civil and military side in the immediate post-war
12

years, and a former head of the DSIR, clearly did not believe in the ‘linear model’. In his
presidential address to the British Association for the Advancement of Science in 1948, he
asked: “to what then shall we attribute the relative decline [of Britain as a great power]? Shall
we argue that a main cause was that research was on too small a scale?” He preferred other
reasons, noting that Sweden and Switzerland had strong technology, but no great strength in
research. His view was that “it is not the general expansion of research in this country that is
of first importance for the restoration of its industrial health, and certainly not the expansion
of government research remote from the everyday problems of industry. What is of first
importance is to apply what is already known.”lx That view may not be incompatible with
elements of the standard ‘linear model’, but it is different from and richer than the whole
linear model.
Putting the Linear Model into the History of Science Policy
There is a suggestion in some of the more recent literature that something very close to ‘the
linear model’ was the core idea in ‘science and technology policy’ after 1945, at least into the
late 1960s. Many historical studies of US and British science policy, even if they do not use
the term ‘linear model’, refer to ‘articles of faith’ and ‘paradigms’ amongst which was
support for basic research on economic grounds, going beyond a ‘science policy elite.’lxi
Bruce Smith, writing of US policy notes that: “While a healthy basic research effort as the
lynchpin of the system was a primary article of faith, the consensus was broad enough to
include those who wanted more basic research, and those who had doubts, because there was
money for all.”lxii In the non-academic literature, the ‘linear model’ looms large as the core of
post-war policy, with implicit and explicit reference to Vannevar Bush. The first example
refers to Britain:
science was seen as the engine of progress and as such worthy of State patronage. The
science policy debate therefore focused on the resource inputs into science, in the
belief that if a country had a sufficient investment in basic science then technological
innovation, economic growth and social progress would surely follow. Within this
‘science push’ phase, the chief policy issues concerned the funding of ‘big’ science,
above all nuclear research; the search was for criteria for choice. Below this strategic
level (in which scientists themselves played a considerable part), the chief mechanism
for resource allocation was peer review. Phase 1 took a fundamentally optimistic view
of science as a quest—an endless frontier. Serendipity would take care of the rest. This
13

linear model of science pushing technology, which policy-makers once saw as a self-
evident truth, has long since gone as the relationship between basic science and
technological innovation has come to be understood as highly complex and quite
difficult to influence.lxiii
The second example is a US embassy summary of a Chinese document on science and
technology policy, which testifies to the ubiquity of the model:
Since 1945 the United States has followed the basic research to applied research to
product development linear model propounded by President Roosevelt’s science
advisor Vannevar Bush in his 1945 book ‘Science—the Endless Frontier.’ According
to this view, which has been fundamental to U.S. S&T policy, basic and applied
research are distinct and not complementary.lxiv
If the content of Science: The endless frontier is misrepresented so its influence is
exaggerated. As historians have long pointed out, its main recommendations were ignored.
Basic research was increasingly funded by government, but neither by the agencies or in the
spirit proposed by Bush. For while Bush proposed the support of basic science across a wide
field by a new agency, many different agencies were to do so, above all the Office of Naval
Research, the Atomic Energy Commission, the National Institute(s) of Health, and so on.
When the National Science Foundation was formed in 1950, it was an addition not a
substitute: the ONR dominated basic research into the early 1950s, even in the early 1960s the
NSF supported less than 10% of all ‘basic research,’ and less than 15% of all federally-funded
‘basic research.’lxv The Department of Defense was still supporting as much as 44% of
federally funded basic research in universities and colleges in 1958.lxvi Did these funding
agencies really believe in free untrammeled university research directed only by curiosity, as
the ‘linear model’ analyses of post-war science policy would imply? Historians are clear: in
the case of physics the military did not believe they were funding generic ‘fundamental
science’ even if the recipients often claimed this.lxvii The ONR did not operate peer review as
its main allocation process—it relied on ‘program managers.’lxviii Forman notes that the
military spent 5% of the military R&D budget on ‘basic’ research, not because this was what
was needed to feed technical development, but because it was a convenient proportion.lxix
Furthermore, “In truth, only a small fraction of that 5% of R&D funds labeled basic research
went to support investigation that could reasonably be called fundamental.”lxx By which he
14

means much was devoted to techniques and applications. It was “a physics as the military
funding agencies would have wished.”lxxi Academic physics was not in command, it was being
used, and they “had lost control of their discipline.”lxxii
‘The linear model’ never dominated innovation policy; nor did it dominate narrower ‘science
policy.’ Academic historians and other analysts have, by granting so much attention to policy
for ‘basic’ or ‘fundamental’ research, reproduced the focus of the ‘linear model’ they criticize,
and shared its assumption that what really mattered was this high level stuff. The rest is
merely derivative. The problem is that the great bulk of research and development (which is
better described as development and research) was not ‘basic’ or ‘fundamental’ or thought of
in terms of the linear model at all: it was driven by quite different concerns. At best the
‘fundamentalists’, to call them something, were arguing for a small space for fundamental
research in a world of applied, directed, and controlled research. They had to argue against
large constituencies of technical experts and researchers and ‘users,’ who demanded control
and direction of research and development, and did indeed control most research and
development. It may well be the case that the ‘fundamentalists’ succeeded, relatively, in
increasing the proportion of ‘fundamental’ or ‘basic’ research in total ‘research and
development,’ but not that fundamental research ever dominated. And yet that is what much
commentary on postwar research and development and innovation policy implies.
Histories of Twentieth Century Science
Most histories of twentieth century science are concerned with the academic basic science that
Bush wanted to promote, and the agencies that did indeed promote them: that, or something
like it, is what ‘science’ is taken to mean. Certainly few general accounts of nineteenth and
twentieth century science, technology and medicine give anything like enough weight to non-
basic, non-academic research. This is true even when the small scale of the academic research
enterprise, compared to ‘research and development’ or even ‘science and engineering’ is
recognized, as in the case of the work of Bruno Latour.lxxiii This bias is obvious in the
importance given to accounts of university research, but where its invisible power is best seen
is in accounts of science and the military, science and government and science and industry.
There is a systematic bias in these accounts towards the study of academic scientists in
relation to these bodies. For the case of Britain, which I know best, there was a genre of
writing on the history of science policy, which systematically confused the history of ‘science
policy’ with the history of academic and related research. For example, the ‘Haldane
15

principle’ of autonomy of research councils, which only applied to ‘fundamental’ research,
and some cross-cutting research, is assumed to be the central guiding principle behind all state
research funding.lxxiv For the US case there is a distinct tendency to focus on the OSRD and
the NSF that it is as if these bodies were the only ones which mattered in mid-century science
policy. In the case of science and the military the story in most literature is one of the entry of
academics scientists into association with the military in times of war, and the continued
funding of academic science (in the US case) by the military in the cold war.lxxv The
historians of technology, have tended to focus on the most ‘academic’ of industrial research.
An account by Donald Cardwell in 1957 made an interesting distinction between technology
and ‘applied science’—technology the application of given laws was ‘the application of the
results of science’ while ‘applied science’ was the actual investigation by the methods of
‘pure’ science, of laws relevant to the industry concerned: it was science restricted to the
‘foreseeable interests of industry.’ Cardwell saw it as new, and explicitly denied that it arose
from other older industrial practices.lxxvi He denied too that the technologists evolved into the
applied scientists—the pioneering German industrial labs were derived from academic
examples, the staff were the products of universities and colleges that otherwise turned out
teachers. A whole generation of studies of industrial research (and I too am guilty) have
focused on research in industry and above all on central corporate research laboratories.
This profoundly academic-research-oriented model of twentieth-century science is all the
more surprising in view of the long tradition of stressing the non-academic origins of modern
science, particularly the craft traditions, and the insistence of much history of science,
strengthened in the last 20 years, on the significance of industrial contexts for science, from
dyeing to brewing to engine making. The idea that academic science is strongly dependent on,
affected by, derivative of ‘technology’ has long been a commonplace of the history of
nineteenth and twentieth century science. In that sense we have long since moved from a
scientific conception of technology to a ‘technological conception of science.’lxxvii Indeed one
historian suggests shifting from the usual historical use of ‘science-based industry’ to
‘industry-based science.’lxxviii Nevertheless there is a strong tendency to look at the industrial,
technological and other contexts of academic science—rather than non-academic science as
such.lxxix
16

Alternatives to the Academic Research Model of History of Science
Bush’s account of the development of research in the United States, is, paradoxically enough,
a good place to start a sketch of an alternative picture.lxxx For Bush pointed out that the great
bulk of research was applied research and was done in government and industry; he
highlighted the increasing proportion of industry and government research. Simply
recognizing the comparative scale of industrial and government research through the century
is enough to transform the usual implicit maps of the twentieth-century research enterprise
historians have worked with.lxxxi The history of innovation and innovation policy must surely
focus on industry and government agencies concerned with it, rather than ‘science policy’.
Such a non-academic perspective can change our accounts of standard linear model cases.
Take for example histories of the atomic bomb project which usually take the form implied by
the linear model—they start with academic physics, and go through ‘big science’ to the
atomic bomb: this is history taken from the biographies of academic physics as they move
from pure to applied. For the historian of technology and the military there are many
precedents, both industrial and military, to the bomb project, as is clear in Thomas Hughes’
quite distinct account of the Manhattan Project.lxxxii Indeed the Manhattan project is seen as
possible because there were such precedents and capacities. But, one could see it a part of a
process of development of the innovative capacity of the military and of large corporations,
which extended their range to ‘pure’ nuclear physics. Indeed the whole of wartime R&D
activity is best seen in this way—as an extension and strengthening of pre-existing military
and industrial organizations, rather than the export of academic basic science into wartime
bodies. Even in the post-war years, when the prestige of basic science, and particularly
academic basic science, was very high, it is still useful to see industrial and military research
as an upgrading of industrial facilities as much as an importation of academic models and
personnel. To be sure, there was a wave of laboratory building far from production, and the
bringing in of high-level academics, but there was expansion in all kinds of scientific activity
after the war.lxxxiii At very best the ‘linear model’ to the extent we grant its existence at all will
be a very small part of the picture.
I want to argue for a more general critique based on the observation that there is a systematic
confusion in the literature between ‘science’ and scientific research, which is hardly noticed
because so many analysts assume science to be research.lxxxiv It is important to distinguish
between expertise, including scientific expertise, and research. The twentieth century belief
that “Science implies the breaking of new ground,”lxxxv has made the history of science and
17

expertise the history of research. Intellectuals, even engineers like Vannevar Bush, generally
used ‘science’ and ‘scientific’ refer to scientific research. This is one reason why the history
of science in business, or of science in government, or the university is, without this being
clear, the history of research.lxxxvi Yet research was something new which increasingly came to
define science, and scientific, even though only a small proportion of ‘scientists’ were ever
engaged in it; the research revolution in the late nineteenth century, not merely a laboratory
revolution, and that it extended right across the field of knowledge, and to many kinds of
institution, was of huge significance.lxxxvii On the other hand we should not ignore the
continuing importance of non-research technical expertise.lxxxviii Nor should we ignore the
historical processes by which ‘research’ became so important, and we should resist the
temptation to see ‘research’ as itself created with the realms of ‘pure science.’ We would do
well to think about twentieth century ‘science’ as a great mass of non-research science, some
‘applied science,’ and a little bit of ‘basic science,’ if we grant the categories. ‘Research’ is not
the norm for scientific activity, and within ‘research’ pure research is not the norm either.
Of course some historians of technology, and some historians of science, have looked at non-
research science and technology, though both subjects are generally innovation and research
centered.lxxxix As Ernst Homburg has pointed out, it is highly misleading to identify the
beginnings of science in industry, with research, or more generally the significance of science
in industry with the significance of research.xc However, in the literature on science in
industry, although dominated by research, scholars have been careful to point to (though often
without giving too much detail) that research was generally built on top of, or out of
significant pre-existing scientific organizations.xci Scientists were first employed in routine
jobs, for example in industry scientists are first employed for production, and the scientific
role is upgraded with time till some is involved in ‘pure science.’ Indeed we can follow a line
of analytical labs, development labs, and research labs, generally established in that order. In
this alternative picture—the actual placing of ‘pure,’ ‘fundamental’ and ‘basic’ research is best
understood as emerging (sometimes) from what academic researchers see as lower kinds of
scientific activity. What we see is not the rise of science as such, but the rising prestige of the
‘researcher.’ This is not to say that particular individuals evolved in the same way —in many
cases, each higher level meant the recruitment of new kinds of personnel from outside
industry.xcii Such developments are familiar in cases of IG Farben, GE, AT&T and Du Pont,
where research laboratories came after the establishment of many other sorts of laboratories.
One of the best-studied cases in this context is IG Farben: it had around 2,000 chemists in the
18

late 1920s, with around 1,000 classified as research chemists. The latter they worked in 50
laboratories, of which ten were large ‘main laboratories.’xciii In Britain too, the pattern was
very similar—with research emerging out of existing scientific activity.xciv In the armed forces
too ‘research’ was added to pre-existing scientific work, but of the uniformed branches and of
civilian technical specialists. Certainly in the interwar years there emerged a large civilian
research corps was gaining in power and prestige in the interwar years, gaining ground over
other technical specialists, later supplemented by academic researchers.
Can a similar story be told for academic research? It can, for most academics were not
researchers until well into the twentieth century. In the case of laboratories, these too were
concerned with teaching when first introduced in the nineteenth century, then a place for
analysis and testing, and then a ‘research laboratory’ focused on a research program.xcv Indeed
the study of industrial research laboratories has provided materials for the reassessment of the
history of academic research. Dennis argues research developed simultaneously, and in
parallel ways in industry and the university.xcvi In medical schools too, laboratories are first
teaching places, and their teaching staff lowly figures compared with the clinicians. Later
research laboratories and researchers clearly have higher relative prestige.xcvii The research
revolution in the universities, as in industry and government, was a slow one: even in the
1930s research was not universal in university departments of arts or sciences.xcviii In short
the ‘linear model’ does not work for institutions, just as it does not work for innovations. The
new institutions of science were not pioneered through fundamental work in the academy and
then progressed linearly down to everyday practice; the reverse might be better, crude,
approximation.
Non-Cumulation
Why such a straw man as ‘the linear model’ gained such currency and such significance in the
academic literature of the 1980s and 1990s is not a question readily answered. Indeed what
needs explanation is not just this case for there was a general tendency in the science and
technology studies literature, and the associated historical literature too, to deploy a number of
other straw men like ‘technological determinism’ and ‘whig history’. One can understand why
this was done—it was convenient to invent labels for naïve positions influentially peddled in
the public sphere by scientists and engineers, and found in particular in the views science and
engineering undergraduates taught by STS and other academics. The very failure of STS to
define the public discourse around science and technology doomed us to attacking public
19

science. Instead of engaging with the arguments of critics of naïve positions, of which there
were by the 1990s several generations, the naïve conceptions still had to be attacked. These
naïve conceptions created by academics thus achieved greater prominence than they would
otherwise have had.xcix The result was a non-cumulation of the critical positions, to the extent
that it came to be argued that even the older academic literature was contaminated. The ‘linear
model’ case reveals how some academics systematically ignored several generations of
academic work on innovation, which presented much richer accounts of innovation than those
criticized. But it is hardly the only case. Let me take some examples from my own work. For
example, the radical British Ministry of Technology of the late 1960s, famous then and now,
did many things, which many analysts hold did not, and could not have happened in Britain.c
Studies of the relations of science and the military and Britain have not maintained the
knowledge present in Bernal’s otherwise famous Social Function of Science.ci Histories of
British industrial research done in the 1970s refuted the key conclusions of well-known work
produced in the 1980s.cii The SCOT program relied to a significant degree on rubbing out
previous generations of studies of ‘social construction.’ciii Perhaps the most devastating
criticism of the linear model—that there is no correlation between national R&D spending
and national economic performance, appears to be unknown to most students of innovation,
certainly most critics of ‘the linear model,’ but was a commonplace in the 1960s.civ For the
history of British science we have had generations of what I have called ‘anti-histories.’cv
Readers will doubtless have their own examples.
In this particular paper I have argued that we should take on board what we have known (in
principle) for several academic generations that the study of industrial innovation and science
in industry, rather than starting yet again with an attack on a straw man. In studying science in
industry we should start with the literature on industry, not academic science, and in particular
from literature that it not driven by academic research model assumptions.cvi What deserves
criticism is our own academic work, not straw men, or models popularized for propaganda
purposes by the academic researchers.cvii The implications are that we should reject the
academic-research-centered model of science, and indeed the research-centered model of
science, which remain dominant, if we are to understand the relations of science and industry
in the twentieth century. While the study of academic research science is interesting in its own
right, it can’t stand for the study of science as a whole, unless, that is, we believe in the ‘linear
model’.
20

21

i I am grateful to the participants at the Symposium for their many and varied comments, which
have helped strengthen the argument of the paper. I am especially grateful to Mats Fridlund,
Andrew Mendelsohn and participants in a doctoral seminar at Imperial College for their invaluable
criticisms of earlier versions. Thanks are also due to Mats Fridlund for many examples of the use of
the term I would otherwise not have seen. I am also grateful to Eric Schatzberg for his observations,
and for some material I would not otherwise have come across.
ii As was clear in the preliminary paper for this Nobel Symposium, and indeed in many of the
abstracts submitted.
iii I have sought to clarify meaning of ‘technological determinism’ in David Edgerton, “De
l’innovation aux usages: Dix theses sur l’histoire des techniques,” Annales HSS, no. 4–5 (Juillet–
Octobre 1998), pp. 815–837. English version, “From Innovation to Use: Ten (eclectic) theses on the
history of technology,” History and Technology, vol. 16 (1999): 1–26. I deal with the particular
inflexions of Whig history in the British case in England and the Aeroplane: An Essay on a Militant
and Technological Nation (London: Macmillan, 1991) and Science, Technology and the British
Industrial ‘Decline’ ca. 1870–1970 (Cambridge: CUP/Economic History Society, 1996).
iv Here I am following on from a paper by Michael Dennis which is insufficiently known among
historians of science (and technology) Michael Dennis, “Accounting for Research: New histories of
corporate laboratories and the social history of American science,” Social Studies of Science, vol. 17
(1987): 479–514.
v I think it is clear from most of the criticisms of the linear model that what is objected to, is not just
the linear sequence of steps, but also the alleged main source of innovation. I do not think it sensible
to associate the term linear model with the argument that basic science was important, or the
separate argument that research managers, say, plotted sequential models. One of the clearest linear
models attacked is that in Terence Kealey, The Economic Laws of Scientific Research (London:
Macmillan, 1996), which takes the model as one starting with publicly-funded academic research.
vi Donald Stokes, Pasteur’s Quadrant: Basic science and technological innovation (Washington
DC: Brookings, 1997), pp. 10, 18-19.

vii Harvey Brooks, “Lessons of History: Successive challenges to science policy,” in The Research
System in Transition, eds. S. Cozzens, P. Healey, A. Rip and J. Ziman (Dordrecht: Kluwer, 1990), p.
13. Kealey, Economic Laws also clearly extends the meaning of ‘linear model’ to include economic
growth.
viii Bruno Latour’s ‘diffusion model’ is in many ways the linear model in another, funnier, guise.
Bruno Latour, Science in Action: How to follow scientists and engineers through society
(Cambridge, MA: Harvard University Press, 1987), pp. 132–144.
ix David E. H. Edgerton, “Research, Development and Competitiveness,” in The Future of UK
Industrial Competitiveness and the role of Industrial Policy, ed. K. Hughes (London: Policy Studies
Institute, 1994), p. 48. I went to state in a footnote that it was “largely a straw man invoked to
demonstrate the superiority of current ways of discussing innovation […] If it existed at all, [it] was
merely an apologia for the funding of pure science,” pp. 53–4, the argument I am developing here.
x Nick Henry, Doreen Massey and David Wield, “Along the Road: R&D, Society and Space”,
Research Policy, vol. 24 (1995), p. 708. The article finds that ‘the linear model’ structures R&D and
its relation to production in the companies they study.
xi Chris Freeman, “The Greening of Technology and models of innovation”, Technological
Forecasting and Social Change, vol. 53 (1996), p. 27.
xii Freeman, ‘The Greening of Technology’, p. 27.
xiii Ernest Braun, Futile Progress: Technology’s Empty Promise (London: Earthscan, 1995), p. 53.
xiv Doreen Massey, Paul Quintas and David Wield, High-Tech fantasies: Science parks in society,
science and space (London: Routledge, 1992), chapter three. Thanks to Mats Fridlund.
xv Roy MacLeod, “Toward a New Synthesis: Chemists and chemical industry in Europe,” Isis, vol.
94 (2003), p. 114.
xvi William J. Price and Lawrence W. Bass, “Scientific research and the innovative process: The
dialogue between science and technology plays an important, but usually nonlinear role in
innovation,” Science, vol. 164 (1969): 802–3.

xvii The latter argument was extremely important in Schumpeter’s account of capitalism. See my
brief account in Industrial Innovation and Research in Business (Cheltenham: Edward Elgar, 1996);
John Jewkes, David Sawers and Richard Stillerman, The Sources of Invention (London: Macmillan,
1958).
xviii J. Langrish et al., Wealth from Knowledge (London: Macmillan, 1972), pp. 72–3.
xix Wealth from Knowledge pp. 33–5.
xx Edwin Layton, “Conditions of Technological Development” in Ina Spiegel-Roesing and Derek de
Solla Price, eds., Science, Technology and Society: A cross-disciplinary perspective (London: Sage,
1977), p. 204. This is the closest I could find to the linear model anywhere in this comprehensive
landmark text, except in a passing reference (p. 234) by Chris Freeman to Layton’s chapter.
xxi Roy Rothwell and W. Zegveld, Reindustrialisation and Technology (London: Longman, 1985),
p. 49.
xxii Nathan Rosenberg, Perspectives on Technology (Cambridge: Cambridge University Press,
1976).
xxiii Chris Freeman, The economics of industrial innovation, Second edition (London: Pinter, 1982),
pp. 194, 196–200. I have not checked the first edition.
xxiv See for example the particularly relevant cases of Karl Kreilcamp, “Hindsight and the Real
World of Science Policy,” Science Studies, vol. 1 (1971): 43–66, and his “Towards a theory of
Science Policy,” Science Studies, vol. 3 (1973): 3–29, and Harold Orlans, ”D&R allocation in the
United States,” Science Studies, vol. 3 (1973): 119–159.
xxv R. R. Nelson and S. G. Winter, “In search of a useful theory of innovation,” Research Policy,
vol. 6 (1977): 36–76.
xxvi Vivien Walsh “Invention and innovation in the chemical industry: Demand-pull or discovery-
push?” Research Policy, vol. 13 (1984): 211–34.
xxvii David C. Mowery and Nathan Rosenberg, Technology and the pursuit of economic growth
(Cambridge: Cambridge University Press, 1989). Though it criticizes the neo-classical economic

approach, which they say does radically separate between basic research, where the key steps are
seen to reside, and the appropriation stage (pp. 4 and 6).
xxviii Rudi Volti, Society and Technological Change, Second edition (New York: St Martin’s Press,
1992).
xxix Bruce L. R. Smith, American Science Policy since World War Two (Washington, DC:
Brookings Institution, 1990). For Britain see Gummett, Scientists in Whitehall (Manchester:
Manchester University Press, 1980).
xxx G. Dosi, “Sources, procedures and microeconomic effects of innovation,” Journal of Economic
Literature, vol. 26 (1988): 1120–1171; R. R. Nelson and Gavin Wright, “The rise and fall of
American technological leadership: The postwar era in historical perspective,” Journal of Economic
Literature, vol. 30 (1992), 1931–1965; Paula E. Stephan, “The economics of science,” Journal of
Economic Literature, vol. 34 (1996), Issue 3, 1199-1235.
xxxi Stuart Macdonald, “Technology beyond machines” in The Trouble with technology:
Exploration in the process of technological change, eds. Stuart Macdonald, et al. (London: Pinter,
1983), p. 31. Thanks to Mats Fridlund.
xxxii Kline, S. J., “Innovation is not a Linear Process,” Research Management, 28:4 (July–August
1985), p. 36. On the latter point see p. 44 also.
xxxiii Steven Yearley, Science, Technology and Social Change (London: Unwin Hyman, 1988), p.
115, citing J. Ronayne, Science in Government (London: Edward Arnold, 1984), p. 44.
xxxiv Trevor Pinch and Wiebe Bijker, “The Social Construction of Facts and Artifacts,” in The
Social Construction of Technological Systems, Wiebe Bijker, Thomas Hughes and Trevor Pinch eds.
(Cambridge, MA: MIT Press, 1987), pp. 22 and 28.
xxxv Arie Rip, “Science and Technology as Dancing Partners,” in Peter Kroes and Martijn Bakker,
Technological Development and Science in the Industrial Age: New perspectives on the Science-
Technology Relationship (London: Kluwer, 1992), p. 233.

xxxvi C. F. Carter and B. R. Williams, Industry and Technical Progress: Factors governing the
speed of the application of science (London: Oxford University Press, 1957); Investment in
Innovation (London: Oxford University Press, 1958) and Science in Industry: Policy for progress
(London: Oxford University Press, 1959).
xxxvii Carter and Williams, Industry and Technical Progress, p. 54.
xxxviii Ibid., p. 56.
xxxix John Jewkes, David Sawers and Richard Stillerman, The Sources of Invention (London:
Macmillan, 1958), pp. 6–7. The authors clearly wanted to make analytical distinctions between
science, invention and development.
xl See my Science, Technology and the British Industrial ‘Decline’ ca. 1870–1970 (Cambridge:
CUP/Economic History Society, 1996); “The ‘White Heat’ revisited: British government and
technology in the 1960s,” Twentieth Century British History (1996) and Terence Kealey, The
Economic Laws of Scientific Research (London: Macmillan, 1996). As an example see John Jewkes,
David Sawers and Richard Stillerman, The Sources of Invention, Second edition (London:
Macmillan, 1969), chapter X: “The last ten years in retrospect.”
xli Gerhard Rosegger, The Economics of Production and Innovation: An industrial perspective,
Second edition (Oxford: Pergamon Press, 1986) [first edition 1980] notes that economists and social
scientists, in order to simplify a complex process of innovation, have used ‘sequential’ or ‘stage’
models, without specifying any, but giving a diagram, which is considerably more complex than the
usual linear model one. He notes, “the stage model provides a very useful framework for the study
of innovative activity” (p. 9). He notes three shortcomings immediately: 1) it involves arbitrary
definition into phases 2) it is unidirectional with no feedback 3) it is useful only for major, visible
innovations (p. 10).
xlii George Wise, “Science and Technology,” Osiris 2nd series (1985): 229–246.
xliii George Wise, “Science and Technology,” Osiris 2nd series (1985): 229–246. There is a whole
literature surveying accounts of science technology relations, which is relevant. See for example:

Alex Keller, “Has Science Created Technology?,” Minerva, vol. 22 (1984): 160–182; R. Kline,
“Construing ‘technology’ as ‘applied science:’ Public rhetoric of scientists and engineers in the
United States 1880–1945,” Isis, vol. 86 (1995): 194–221.
xliv Sir Edward Appleton, “Fundamental research and industrial progress,” in Federation of British
Industries, Industry and Research (London: Pitman, 1946), p. 14.
xlv Albert H. Rubinstein ed., Coordination, Control and Financing of Industrial Research:
Proceedings of the fifth annual conference on industrial research, June 1954, with selected papers
from the fourth conference, June 1953 (New York: King’s Crown Press, Columbia University,
1955).
xlvi F.B. Tuck, Ideas, Inertia and Achievement: a survey of current opinion on how to shorten the
time lag between scientific discovery and engineering application. New York: American Society of
Mechanical Engineers, 1960. Thanks to Eric Schatzberg
xlvii Freeman, “The Greening of Technology,” p. 27.
xlviii Donald Stokes, Pasteur’s Quadrant: Basic science and technological innovation (Washington
DC: Brookings, 1997).
xlix Ibid., p. 3.
l Ibid., p. 10.
li Ibid., pp. 18–19.
lii Science, The Endless Frontier: A Report to the President by Vannevar Bush, Director of the
Office of Scientific Research and Development, July 1945 (United States Government Printing
Office, Washington: 1945). I don’t know of any work that analyses the argument of the report with
the exception of part of Ronald Kline, “Construing ‘technology’ as ‘applied science:’ Public rhetoric
of scientists and engineers in the United States 1880–1945”, Isis, vol. 86 (1995): 194–221, but the
argument here goes further. For the political background see: Daniel Kevles, “The National Science
Foundation and the debate over postwar research policy 1942–1945: A political interpretation of
Science—The Endless Frontier,” Isis, vol. 68 (1977): 5– 26; Jessica Wang, American Science in an
Age of Anxiety: Scientists, Anticommunism & the Cold War (University of North Carolina Press,
1999); David Hart, Forged Consensus: Science, technology and economic policy in the United
States, 1921–1953 (Princeton: Princeton University Press, 1997)
liii Paul Forman, “Behind Quantum Electronics: National security as a basis for physical research in
the United States, 1940–1960,” HSPS, vol. 18 (1987), p. 152.
liv It is claimed by Stokes that Bush coined the term ‘basic research;’ but Benoit Godin, “Measuring
Science: Is there ‘Basic Research’ without statistics”, (Mimeo, Montreal, Observatoire des Sciences

et des Techniques, 2000) shows it was used by Julian Huxley in Scientific Research and Social
Needs (London: Watts, 1934). Thanks to Mats Fridlund.
lv The key statistical background was the following: “In the decade from 1930 to 1940 expenditures
for industrial research increased from $116,000,000 to $240,000,000 and those for scientific
research in Government rose from $24,000,000 to $69,000,000. During the same period
expenditures for scientific research in the colleges and universities increased from $20,000,000 to
$31,000,000, while those in the endowed research institutes declined from $5,200,000 to
$4,500,000.”
lvi For the importance of distinguishing ‘research’ from the stock of ‘knowledge’ see for example, S.
J. Kline, “Innovation is not a Linear Process,” Research Management, 28:4 (July–August 1985), p.
36 and Keith Pavitt’s paper in this volume.
lvii Arie Rip, “Implementation and evaluation of science and technology priorities and programs,” in
S. Cozzens, et al., Research System in Transition, p. 274.
lviii Edward R. Weidlein and William A. Hamor, Science in Action: a sketch of the value of
scientific research in American industries (New York: McGraw Hill, 1931), p. 278. Thanks to Eric
Schatzberg.
lix Ibid., p. 267.
lx Sir Henry Tizard, “The passing world,” Presidential Address BAAS, September 1948.
lxi Smith, American Science Policy, p. 78.
lxii Ibid., p. 40.
lxiii From SET Forum Shaping the Future: A policy for science engineering and technology (1995)
<http://www.shef.ac.uk/~sfl/textonly/policy/set03.html>.
lxiv Science and Education for a Prosperous China: Lessons From Abroad: A report from U.S.
Embassy Beijing November 1996 <http://www.fas.org/nuke/guide/china/doctrine/stabrd4.htm>.
lxv Bruce L. R. Smith, American Science Policy since World War Two (Washington, DC: Brookings
Institution, 1990), p. 51.
lxvi Harvey M. Sapolsky, Science for the Navy: A history of the office of Naval Research (Princeton:
Princeton University Press, 1990), Table A-5, p. 137.

lxvii That the military didn’t believe it is argued by Paul Forman, in for example, his “Into Quantum
Electronics” in National Military Establishments and the Advancement of Science and Technology,
eds. Paul Forman and José Manuel Sánchez-Ron (Dordrecht: Kluwer, 1996), pp. 270–271.
Sapolsky, Science and the Navy, p. 63 notes the use by the ONR of a ‘two-title’ policy, whereby
scientists used one title for the project, while ONR staff used another in their relations with naval
staff, the latter stressing naval programmes and needs.
lxviii Sapolsky, Science for the Navy, pp. 94–6.
lxix Paul Forman, “Behind Quantum Electronics: National security as a basis for physical research
in the United States, 1940–1960,” HSPS, vol. 18 (1987), pp 198–9.
lxx Ibid.,p. 216.
lxxi Ibid., p. 224.
lxxii Ibid., p. 229.
lxxiii For the recognition of the small scale of what ‘the diffusion model’ or indeed ‘the linear
model’ takes to be science, see Bruno Latour, Science in Action: How to follow scientists and
engineers through society (Cambridge, MA: Harvard University Press, 1987), pp. 162–73.
lxxiv Tom Wilkie, British Science and Politics since 1945 (Oxford: Blackwell, 1991) while
recognising Haldane’s distinction, is overwhelmingly about the types of research associated with
the research councils. See also Jon Agar, Science and Spectacle (Amsterdam: Harwood Academic,
1998), pp. 2–5; though note that the evidence provided negates the central point.
lxxv This is the burden of Forman, “Behind Quantum Electronics” and Stuart W. Leslie, The Cold
War and American Science: The military-industrial-academic complex at MIT and Stanford (New
York: Columbia University Press, 1993), which recognises in the first few pages that most R&D
was not academic. See also See also E. Mendelsohn, M. R. Smith, & P. Weingart, eds., Science,
technology and the military, 2 volumes (Dordrecht: Kluwer, 1988); the journal Historical Studies in
the Physical and Biological Sciences, for the 1990s in particular, and the special issue on Science in
the Cold War, of Social Studies of Science, vol. 31 (2001): 163–197; Michael Aaron Dennis, “‘Our

First Line of Defence:’ Two University Laboratories in the Postwar American State,” Isis, vol. 85
(1994), pp. 427–55; Everett Mendelsohn, “Science, scientists and the military,” in Science in the
Twentieth Century, eds. J. Krige and D. Pestre (London: Harwood Academic, 1997); A. Pickering,
“Cyborg History and the World War II Regime,” Perspectives on Science, 3 (1995): 1–48.
lxxvi D. S. L. Cardwell, The Organisation of Science in England: A retrospect (London: Heinemann,
1957), pp. 10–11.
lxxvii Rachel Laudan, “Natural Alliance or forced marriage—changing relations between the
histories of science and technology,” Technology and Culture, vol. 36 (1995): S17–S28.
lxxviii Wolfgang Koenig, “Science-Based Industry or Industry Based Science? Electrical
Engineering in Germany before World War 1,” Technology and Culture, vol. 37 (1996): 70–101,
clearly prefers the latter.
lxxix Jean-Paul Gaudillière and Ilana Lowy, The invisible industrialist: Manufactures and the
production of scientific knowledge (London: Macmillan, 1998); Soraya Boudia and Xavier Roqué,
“Science, Medicine and Industry: The Curie and Joliot-Curie Laboratories,” History and
Technology (1997).
lxxx Pickstone’s model of different kinds of ways of knowing which appear in a particular historical
sequence, but adding to and interacting with earlier ones, without replacing them suggests both that
it may have some warrant and some John V. Pickstone, Ways of knowing: A new history of science,
technology and medicine (Manchester: Manchester University Press, 2000/Chicago: Chicago
University Press, 2001). He labels them analytical, experimental and techno-scientific ways of
knowing, though the last is also a way of making. These ways of knowing are found roughly
contemporaneously in science, technology and medicine, and also, though this point is not explored
in any detail in Pickstone’s work in different kinds of institutions: academic, industrial and
governmental. Pickstone’s model does not fully describe the process for the twentieth century, and
does not use the categories above.

lxxxi See my England and the Aeroplane (London: Macmillan, 1991) and Science, Technology and
the British Industrial Decline (Cambridge: CUP/Economic History Society, 1996) for examples of
this.
lxxxii Thomas Parke Hughes, American Genesis (New York: Viking, 1989), chapter eight.
lxxxiii D. E. H. Edgerton, “Science and Technology in British Business History,” Business History,
vol. 29 (1987): 91, 92, 98; Sally Horrocks, “Enthusiasm constrained? British industrial R&D and
the transition from wars to peace, 1942–51,” Business History, vol. 41 (1999): 42–63; D. A.
Hounshell and J. K. Smith, Science and Corporate Strategy: Du Pont R&D (Cambridge:
Cambridge University Press, 1988).
lxxxiv Edgerton, “From Innovation to Use.”
lxxxv Hyman Levy, Modern Science (London: Hamish Hamilton, 1939), p. 710.
lxxxvi See my introduction to Industrial Innovation and Research in Business (Cheltenham: Edward
Elgar, 1996).
lxxxvii See D. Cahan, “The institutional revolution in German physics, 1865–1914”, Historical
Studies in the Physical and Biological Sciences, 15 (1985): 1–65; my “From Innovation to Use;”
John Pickstone, Ways of Knowing (Manchester: Manchester University Press, 2000); Robert Fox
and Anna Guagnini, Laboratories, Workshops and Sites: Concepts and practices of research in
industrial Europe, 1800–1914 (University of California, Berkeley, 1999); Elsbeth Heaman, St
Mary’s: The history of a London Teaching Hospital (Montreal: McGill UP, 2003); Mark Pendleton,
“‘A Place of Teaching and Research:’ University College London and the Origins of the Research
University in Britain, 1890–1914,” University of London PhD, 2001. For the business case see the
papers I edited in Industrial Innovation and Research in Business (Cheltenham: Edward Elgar,
1996).
lxxxviii Or indeed in many other fields, technical and otherwise. Indeed the distinction between
research and non-research can help clarify the debates within interwar British medicine discussed
by Chris Lawrence in his “A tale of two sciences: Bedside and bench in twentieth century Britain,”

Medical History, vol. 43 (1999): 421–49 and “Still incommunicable: Clinical holists and medical
knowledge in interwar Britain,” in Greater than the Parts: Holism in Biomedicine 1920–1950, eds.
C. Lawrence and George Weisz (New York: Oxford University Press, 1998), pp. 94–111.
lxxxix Edgerton, “De l’innovation aux usages.”
xc Ernst Homburg, “The emergence of research laboratories in the dyestuffs industry, 1870–1900,”
British Journal for the History of Science, 25 (1992): 91–111. A point also forcefully made in
Michael Dennis, “Accounting for Research: New histories of corporate laboratories and the social
history of American science,” Social Studies of Science, vol. 17 (1987): 479–518 and especially
Ulrich Marsh, “Strategies for Success: Research organisation in the German chemical companies
until 1936,” History and Technology, vol. 12 (1994), 23–77. See also Pickstone, Ways of Knowing,
p. 171; W. Koenig, “Science-based industry or industry-based science? Electrical engineering in
Germany before World War I,” Technology and Culture, 37 (1996): 70–101.
xci Ulrich Marsh, “Strategies for Success: Research organisation in the German chemical companies
until 1936,” History and Technology, vol. 12 (1994): 23–77, but see also the standard accounts such
as Leonard S. Reich, The Making of American Industrial research: Science and business at GE and
Bell, 1876–1926 (Cambridge: Cambridge University Press, 1985); D. A. Hounshell and J. K. Smith,
Science and Corporate Strategy: Du Pont R&D (Cambridge: Cambridge University Press, 1988)
[note how in these titles ‘science’ is equated with research and R&D].
xcii See Homburg, 109–110 for the German dye firms in the 1880s, which were generally recruiting
their research leaders from the outside.
xciii Ulrich Marsh, “Strategies for Success: Research organisation in the German chemical
companies until 1936,” History and Technology, vol. 12 (1994): 23–77, on p. 23.
xciv David Edgerton, “Industrial Research in the British Photographic Industry 1879-1939,” in ed.
Jonathan Liebenau, The Challenge of New Technology: Innovation in British Business since 1850
(Aldershot: Gower, 1988), pp. 106–134. Sally Horrocks, “Consuming Science: Science, technology
and food in Britain, 1870–1939,” (University of Manchester, PhD Thesis, 1993).

xcv D. Cahan, “The institutional revolution in German physics, 1865–1914,” Historical Studies in
the Physical and Biological Sciences, 15 (1985): 1–65; G. Gooday, “Precision Movement and the
Genesis of Physics Teaching Laboratories in Victorian Britain,” British Journal for the History of
Science, 23 (1990), 25–51; S. Schaffer, “Late Victorian Metrology and its Instrumentation: A
Manufactory of Ohms,” in Invisible connections: Instruments, institutions, and science, eds. Robert
Bud & Susan E. Cozzens (Bellingham, Wash.: SPIE Optical Engineering Press, 1992), p. 23–56.
xcvi Similarly in studies of industrial research there is a sense that the work done even in the most
science oriented of corporate laboratories is to be contrasted with what is taken to be the pure work
of the academic scientist, a criticism forcibly and clearly made by Dennis. Michael Dennis,
“Accounting for Research: New histories of corporate laboratories and the social history of
American science,” Social Studies of Science, vol. 17 (1987): 492ff. Dennis makes the crucial point
that academic research schools were often as narrow in focus as corporate research laboratories (p.
504).
xcvii Elsbeth Heaman, St Mary’s: The history of a London Teaching Hospital (Montreal: McGill UP,
2003).
xcviii For the low place of research in the British university in the interwar years see: Bruce Truscot
(Pseud.) Red Brick University (Harmondsworth: Pelican, 1951), chapter 4. See also Bernal, Social
Function; Mark Pendleton, “‘A Place of Teaching and Research:’ University College London and
the Origins of the Research University in Britain, 1890–1914,” (University of London, PhD 2001).
xcix In a collection of well-known articles on industrial research I deliberately included papers on
the same topic from different periods, not only to show how historical interpretation had changed,
but also in some cases retrogressed. David Edgerton (Ed.), Industrial Research and Innovation in
Business (Cheltenham: Edward Elgar, 1996).
c “The ‘White Heat’ revisited: British government and technology in the 1960s,” Twentieth Century
British History, vol. 7 (1996): 53–82. Reprinted in Luca Guzzetti ed., Science and Power: The
Historical Foundations of Research Policies in Europe (Brussels: European Commission, 2000),

pp. 207–236.
ci “British Scientists and the relations of Science and War in Twentieth Century Britain,” in
National Military Establishments and the Advancement of Science: Studies in Twentieth Century
History, eds. Paul Forman and J. M. Sanchez Ron (Dordrecht: Kluwer, 1996), pp. 1–35.
cii Edgerton “Introduction,” Industrial Research and Innovation in Business (Cheltenham: Edward
Elgar, 1996), pp. x–xvi part of International Library of Critical Readings in Business History. See
the paper by Michael Sanderson (1972), which refuted David Mowery (1986).
ciii Thus the SCOT program was an ‘application’ of sociology of scientific knowledge to
technology, oblivious or dismissive of the point that sociology, economics, technology, had been
long been used in the study of technology. The historical novelty and significance of SSK was that
sociology was, long after it had been applied even to other kinds of knowledge, applied to scientific
knowledge. See my “Tilting at Paper Tigers” an essay review of Donald MacKenzie, Inventing
Accuracy, in British Journal for the History of Science, vol. 26 (1993): 67–75.
civ See my Science, Technology and the British Industrial ‘Decline’ ca. 1870–1970 (Cambridge:
CUP/Economic History Society, 1996), “The ‘White Heat’ revisited: British government and
technology in the 1960s,” Twentieth Century British History (1996) and Terence Kealey, The
Economic Laws of Scientific Research (London: Macmillan, 1996). As an example see John Jewkes,
David Sawers and Richard Stillerman, The Sources of Invention, Second edition (London:
Macmillan, 1969), chapter X: “The last ten years in retrospect.” As evidence that the science policy
community is unaware of/indifferent to this fundamental point, see, for example As evidence that
the science policy community is unaware of/indifferent to this fundamental point, see, for example:
Ben Steil, David G. Victor And Richard R. Nelson (eds), Technological Innovation and economic
performance. A Council for Foreign Relations Book (Princeton: Princeton University Press, 2002).
cv See my forthcoming Warfare State.
cvi See Steven Shapin’s article in this volume.

cvii Paul Forman, “Independence, Not Transcendence, for the Historian of Science,” vol. 82 Isis
(1991), 71–86; “British Scientists and the relations of Science and War in Twentieth Century
Britain,” in National Military Establishments and the Advancement of Science: Studies in Twentieth
Century History, Paul Forman and J.M. Sanchez Ron eds. (Dordrecht: Kluwer, 1996).