Technology, research, development, and science

The quotations on Introduction and Conclusions refer to science, technology and innovation; they take the view that scientific concepts precede technological advances. This page looks at how science and technology are described, and how they operate, according to academic and other studies. The problem arises that different authors have different ideas of what they mean by science and technology. The first section below looks at how these terms are used, and why, in four accessible and influential sources. It becomes evident that the view that science precedes technology is in accord with the headline information in popular sources, but even a slightly deeper reading shows disagreement.

The next section focuses sources of innovation, and looks at some more specialist sources in the many fields that have an interest in the relations of science and technology. These pose questions about how research and development operate, and what goes on in specific research efforts. Some answers are then found from an examination of research statistics and aspects of a recent major EU programme. Questions remain, though, about academic research and its history. The latter is treated on A problem in the history of science and technology and Raising water; the former is the subject of the the third section. The connections between academic research and education are followed up on the Technology and science education page.

Four sources

Quotations from the four sources are shown in the table below. The first two sources are internet encyclopedias, which represent current informed popular opinion. The next is Science‒The Endless Frontier, which is often cited as the launchpad for the science precedes technology view. It is a report from the USA Office of Scientific Research and Development in 1945: The war had brought unprecedented levels of funding and direction to R&D, and this report was in response to presidential questions on how research results could be used and research continued after the war ended. The fourth source is the Frascati Manual, to represent a current expert view. The manual provides Guidelines for Collecting and Reporting Data on Research and Experimental Development for the OECD. The quotations are from Chapter 2, Concepts and definitions for identifying R&D in the 2015 edition.

The table shows quotations from each source that were located by searching for each term in the left hand column.

Britannica and Wikipedia

Anyone who encountered the view that science precedes technology and did a quick fact check in the internet encyclopaedias would find that it checked out. The first things that come up in Britannica and Wikipedia confirm that new knowledge comes only from basic research, and in Britannica that technology is just the application of existing science. But a further look into these sources reveals opposing views in articles relating to technology. Elsewhere in Britannica there is: Technology is the systematic study of techniques for making and doing things; science is the systematic attempt to understand and interpret the world, and in Wikipedia: The exact relations between science and technology in particular have been debated by scientists, historians, and policymakers in the late 20th century, in part because the debate can inform the funding of basic and applied science. … The issue remains contentious, though most analysts resist the model that technology simply is a result of scientific research.

It appears that the views expressed depend on the interests of the authors. Those focused on science present technology as straightforwardly a product of science, and it is their view that has grabbed the headlines. Those specialising in technology see the matter as more complex, and have pointed out the connection with the funding debate. This connection plays an important role in the third source in the table.

The Endless Frontier (EF)

According to EF, knowledge is produced by Basic Research, used by Applied Research, and Technological progress is confined by the output of Basic Research. EF was a campaigning document, and it gave a simple message for the times: that basic research is essential and valuable. One reason for broadcasting this message is given in EF: under the pressure for immediate results, and unless deliberate policies are set up to guard against this, applied research invariably drives out pure. The campaign was to ensure funding for basic research amid the uncertainty of government funding after the war.

Since the time of EF, other interested parties sought to identify links between basic research and technical innovation. One study was Project Hindsight (PirtleProject Hindsight Final Report (Office of the Director of Defence Research and Engineering, Washington, D.C., 1969).) conducted by the US Office of the Director of Defense Research and Engineering. This identified R&D events in the development of weapons systems in the 20 years prior to 1965, and found that only three out of 710 events could be attributed to basic research. It was countered by two studies, Technology in Retrospect and Critical Events in Science (TRACESTechnology in Retrospect and Critical Events in Science, Illinois Institute of Technology Research Institute, (National Science Foundation, Washington D.C., 1968).), and Interaction of Science and Technology in the Innovative Process: Some Case Studies (BattelleBattelle Research Institute, 1973. Interaction of Science and Technology in the Innovative Process: Some Case Studies, Battelle Research Institute, processed ). These looked at ten innovations and the events in their origins over a much longer 50 year period. The former study concluded that basic research provided 70% of the events, and the latter study that the figure was 34%, reducing to 15% for critical events.

This pattern of assertion and refutation continued after a 1976 investigation of the Scientific basis for the support of biomedical science (Comroe & DrippsComroe JH, Dripps RD. Scientific basis for the support of biomedical science. Science 1976;192: 105-11.). They looked at the origins of the top ten clinical advances in cardiovascular and pulmonary medicine and surgery in the last 30 years and concluded that 41% of over 500 key articles that led to these advances were written by scientists who had no interest in disease and that 62% were the result of basic research. Among their assumptions was that any research that sought explanations in underlying causes or models counted as basic research. Subsequent reviews of these findings showed first that the method for obtaining these results was biased (SmithComroe and Dripps revisited, Richard Smith, British Medical Journal Volume 295, 28 November 1987 p1404), and subsequently that the method could not be repeated and the conclusion is not reliable (Mason et alFactors that lead to advances in neonatal intensive care: Comroe and Dripps revisited, Barbara Mason, Elizabeth Green and Jonathan Grant, Research Evaluation, volume 10, number 2, August 2001, pages 121–127).

A reason for these unsatisfactory research outcomes may be that they sought clear and simple links which do not exist. In "The influence of market demand upon innovation: a critical review of some recent empirical studies" Mowery & RosenbergDavid Mowery and Nathan Rosenberg, Research Policy 8 (1979) 102-153 concluded that the simple models of research feeding innovation or vice versa did not hold water: the relationship between technology and science is more complex than that. One of the problems in the longer term studies is that the origins of an innovation are not clearly and objectively defined over this time scale. Gibbons & Johnstonthe roles of science in technological innovation, Michael Gibbons and Ron Johnston, Research Policy 3 (1974) 220-242 found that a better defined result is obtained when the immediate sources of solutions to technical problems are examined. They concluded that they come from close relations between industry and researchers in universities and other institutions. This is the same conclusion reached by the UK Council for Scientific Policy report Third Report, Cmnd.5117 (H.M. Stationery Office, London, 1972) p.32: We cannot see that it is possible in any systematic way to trace important industrial applications of science back to basic work … in a way which could help in determining how much support is justified and we would now place less emphasis on wholesale importation of isolated scientific developments, and more emphasis on a generalized diffusion of knowledge….

It appears that the relationship between science and innovation is not as clear as the campaigners for the funding of basic research would have it be. But then "its complicated" does not make a good campaign slogan. Nonetheless, funding decisions have to be made, and that may involve classifying types of research and their interrelations. This is one aim of the fourth source in the table.

The Frascati Manual

The Frascati Manual acknowledges the complex relationship between basic research and technical development. It states (p45): The order in which the three types of R&D activity appear is not meant to suggest that basic research leads to applied research and then to experimental development. There are many flows of information and knowledge in the R&D system. Applied Research and Experimental Development are described as a source of knowledge, and technology as a source of scientific efforts. The term Science is only used to refer to the activities of various disciplines, and not linked to a distinctive way of doing research. It advises that While the manual has always applied to all scientific disciplines, [in the current edition] there is more emphasis on the social sciences, humanities and the arts, in addition to the natural sciences and engineering. Elsewhere in the manual there is no association of different types of research or development with different disciplines, and no suggestion that any particular set of disciplines is basic or fundamental.

The guidelines given for identifying the types of the research activities involved in a research project say that basic research has a longer lead time to application, and a broader potential field of application, than applied research. However, the example they choose to illustrate the flow of knowledge is that applied research and experimental development could adapt fundamental knowledge arising from basic research and that there would be feedback that takes place when knowledge is used to solve a problem. Although there is no formal statement to that effect, their thinking appears to be that basic research is the source of knowledge which is then used by other types of research.

This thinking appears too when the manual gives numerous examples of how to differentiate its three main types of R&D in practice. The examples in life and physical sciences are all structured in the order basic—applied—development. It appears that among the many flows of information and knowledge in the R&D system the authors had a particular subset in mind when drafting the manual. They do not include occasions when experimental development reveals unanticipated new phenomena subsequently investigated by basic research. Basic Research is further divided into sub-types of Pure and Oriented, the latter intended to form the basis of the solution to recognised or expected current or future problems or possibilities. All their examples of Basic Research appear to be Oriented.

The manual divides R&D activities among four sectors: Business enterprise, Government, Higher education, Private non-profit. All three types of R&D may be involved in a single research project. All three types of R&D may be carried out in any of the sectors, and basic research is usually performed in the Higher education sector. The latter statement is only partially supported by the data on OECD research expenditure by type and sector shown in the table at the right.

The Frascati Manual has abandoned the simple view, but at the time of writing its entry in Wikipedia has not. The entry describes the manual as defining Experimental development is systematic effort, based on existing knowledge from research or practical experience, directed toward creating novel or improved materials, products, devices, processes, systems, or services. This leaves out the production of new knowledge by the most applied type of R&D. It also makes little sense, since if Experimental Development is required to create novel materials, etc then the knowledge of how they behave must be new.

Review

What does all this say about the view that science precedes technology? This view has been advanced in the past as part of an effort to encourage ongoing central funding for academic research. It has not been substantiated by subsequent investigations, and remains the subject of debate. It has produced bias in academic researchers' estimation of their own contribution to the development of new applications. It is no longer the view that governs research policy. It is the view that still appears most prominent in popular sources. The terms science and technology mean different things to different people, so there needs to be some clarification if the debate is to progress.

This throws some light on the ways the view has been expressed in the extracts on Introduction and Conclusions. The striking thing about the statements of Collini, Ferguson and the Britannica article that they show no awareness that they concern matters of debate. They are unreserved and unsubstantiated: they present a truth universally acknowledged. The Britannica article on The Urban Revolution offers no analysis of the historical record to demonstrate that science provided the means for technical progress, but simply states that it is the only conceivable possibility. ColliniWhat are universities for?, Stefan Collini, Penguin 2012, p55 has campaigned for the independence of academic research, and has identified one mark of an academic discipline as being that the open-ended quest for understanding has primacy over any application or immediate outcome. His assumption, that the historical record shows that this quest commonly enables useful outcomes, is deployed in support of his campaign. It should be reiterated here that the focus on this page on technical innovation does not in any way suggest that it should be the only criterion for deciding on what research to undertake. The aim here is to clarify how innovation arises, and not to say anything about what kind or how much of it we should have.

Ferguson's remark on the origins of the steam engine is a travesty of the closely argued views of WoottonDavid Wootton, The invention of science : a new history of the scientific revolution, Penguin, 2015 and others, that the development of a useful steam engine was enabled by prior academic research. This topic plays an iconic role in the historical debate, and is examined in detail on A problem in the history of science and technology and Raising water. The extract from the UK National Curriculum shows how this particular view of how science and technology operate has a significant impact on general education. There are other views among education specialists, and there are other curricula; these are the looked at on Technology and science education.

So far on this page the discussion has been around the popularly prevailing view that science precedes technology, and has demonstrated that all but the most superficial inquiry shows that it is at least inadequate, and may well be seriously mistaken. Nonetheless, a great deal of innovation is achieved, and a great deal of research results are produced. Although there is no simple model for the structures and processes involved in this, there is as always plenty of material to draw on in trying to understand it. The questions of how technology and science operate, and how they have related to each other now and in the past, have engaged the attention of specialists in history, economics, research policy, and science and technology studies, as well as education.

Where does innovation come from?

Economic history

In a review of the historiography of technical progress, Rosenberg writes Clearly, the issues turn, in part, upon definitions and how rigorously one defines science but is also able to say that

What is certainly clear and is borne out by the histories of England, France, the United States, Japan, and Russia over the past two and a half centuries or so is that a top-quality scientific establishment and a high degree of scientific originality have been neither a necessary nor a sufficient condition for technological dynamism.

(RosenbergInside the black box : Technology and economics, Nathan Rosenberg, Cambridge : Cambridge University Press 1983, p13). He also says that given the economic incentives, it is hardly surprising that the normal situation in the past, and to a considerable degree also in the present, is that technological knowledge has preceded scientific knowledge (RosenbergInside the black box : Technology and economics, Nathan Rosenberg, Cambridge : Cambridge University Press 1983, p144). This clearly contradicts the casual assumption that in the past the key to innovation lay with science, but still leaves open the possibility that in more recent times science has played a larger rôle.

History of technology

According to FriedelRobert D Friedel, A culture of improvement: technology and the western millennium, Cambridge, MA : MIT Press, 2007's book on technology and the western millennium, the main source of technological change is improvement, not discovery or invention:

Concepts such as invention and [scientific, industrial] revolution tend to direct our attention to relatively few individuals in giving change its shape and force. But, in fact, most technological change—improvement, if you will—comes through the small contributions of ordinary, anonymous workers and tinkerers.

This is a perspective that should be born in mind when comparing types of research and their outcomes: all types of research only add up to a subsidiary contribution to innovation. Friedel's broad statement introduces a book that covers a thousand years of western history, and was not intended to apply uniformly to all times and all places. The statement remains current, however: in a 2018 conference paper on the early promise of nanotechnology, Richard Jones writes: Maybe our constant focus on revolutionary innovation blinds us to the real achievements of incremental innovation, and also:

In a field like nanotechnology, relatively incremental developments of existing technology coexist with much more radical possibilities, and this leads to a tension: the promise is sold on the grand vision and the big metaphors, but the achievements are largely based on the aspects of the technology with the most continuity with the past.

This shows that David EdgertonDavid Edgerton, The Shock of the Old, Technology and Global History Since 1900, Profile, 2008's portrayal of Technology and Global History Since 1900 as The Shock of the Old also remains current. In a chapter on Invention, Edgerton describes how The university was keeping up with a changing technological world rather than creating it (p186), and That there is a particularly widespread belief in the significance of academic science as a source of invention is testimony to the great influence of academic research scientists (p187). He also points out (p185) that most scientists are not researchers, so most science is not research. These observations can be looked at quantitatively.

How much science and how much research

In the European Union (EU-28), "HRST core" is defined as persons who have completed a tertiary level of education and work as science and engineering professionals, health professionals, or information and communications technology professionals. In 2018 there were 53.7 million persons employed in the EU-28 as HRST core. The EU-28 figures for employment in research show that in 2016 there were 0.95 million working within the business enterprise sector, 0.72 million in the higher education sector, and the government sector employed 0.20 million. They define researchers as professionals engaged in the conception or creation of new knowledge, products, processes, methods and systems and in the management of the projects concerned, so perhaps they include researchers in all fields, and the numbers in science and technology will be somewhat less than these.

These figures show that in terms of employment in science and technology, practice outweighs research by about 30:1, and outweighs academic research by about 75:1. If we are to believe in the dominance of research as a source of innovation, it would have to be argued that the per capita output of the practitioners was less by an enormous factor in quantity, or in impact, than the researchers.

There is an obvious argument for there being a quantity factor, in that the practitioners are restrained from creativity and innovation by their conditions of employment and education. It is doubtful whether this factor could be large enough to reject Friedel's view that improvement is still the main source of technical change. In any case, the larger it is, the more call there is to deploy resources to alleviate the restraints on creativity that enlarge it.

The argument for an impact factor is the one that has been investigated with such unsatisfactory outcomes as discussed above. This argument is also investigated for one case in the history of science on the A problem in the history of science and technology page; in that case, it appears to be unsubstantiated. The argument is accepted without question in the education systems in some countries, and is used in support of curriculum change or stasis.

From the above considerations it appears that gradual improvement may have been the major source of innovation in the past, and may continue to be so in some form in the present. That still leaves out a great deal of detail about how research and practice are related: this is the subject of the next section.

Types of research, and connections to practice

The great majority of research and development has the intention of helping to achieve practical objectives. This is demonstrated in the OECD funding table, in the EU FP7 budget allocations given below, and has been reported by other authors (StokesPasteur's quadrant : basic science and technological innovation, Donald E. Stokes, Washington, D.C. : Brookings Institution Press, 1997, Jones). The research is directed not only to creating and improving industrial products (eg a better battery) but also to improving our understanding of natural phenomena (eg climate being investigated with the objective of predicting it, or disease with the objective of preventing it).

Research policy

In the summary of a 1967 National Academy of Sciences Report to the US House of Representatives, Harvey Brooks wrote

Moreover, the fact that research is of such a nature that it can be applied does not mean that it is not also basic. Almost all of Pasteur’s work, from the fermentation of beet sugar and the disease of silkworms to the anthrax disease of sheep and the cure of rabies, was on quite practical problems; yet it led to the formulation of new biological principles and the destruction of false ones, which revolutionized the conceptual structure of biology.

Donald StokesPasteur's quadrant : basic science and technological innovation, Donald E. Stokes, Washington, D.C. : Brookings Institution Press, 1997 pointed out that the assumption made in the Frascati manual, that applied and basic were mutually exclusive categories, was unjustified: they could be regarded as independent characteristics of research. He set out a quadrant model of research types, according to whether or not they were motivated by use or fundamental understanding, illustrated by famous names with Pasteur's in the yes-yes quadrant (see left). He suggests that the no-no quadrant might be occupied by research motivated by curiosity that catalogues natural phenomena.

Anderson, Herriot and HodgkinsonThe practitioner–researcher divide in Industrial, Work and Organizational (IWO) psychology, Anderson, Herriot and Hodgkinson, Journal of Occupational and Organizational Psychology (2001), 74, 391–411 looked at the work of practitioners and researchers in Industrial, Work and Organizational psychology. They argued that, although relevance and rigor had been regarded as mutually exclusive, both requirements are of crucial importance to our discipline, but that they are not always both met. They presented a simple 2×2 model along the dimensions of relevance and rigour which does not require us to choose between practical relevance and methodological rigour (see right). They report organisational pressures away from rigour, and academic pressures away from practical relevance. They hope that most research in the Puerile quarter is rejected for publication; one of the scientometric studies below indicates that may not generally be the case.

Both these studies overthrow oppositions that have been employed in classifying research into types. Other such oppositions include theory vs experiment (Schauz & Lax p66) and hard vs soft science (Storer).

Harvey Brooks made another point in his summary:

The essential point is that the categorization of research depends on the existing situation in technology and also on the environment in which it is conducted. As definite categories, basic and applied tend to be meaningless, but as positions on a scale within a given environment they probably do have some significance.

The problems with classifying research arise because the categories have been applied too broadly; they are only suitable for restricted use in particular situations. The next section follows that advice.

A particular research and development programme

The European Union's 7th Framework Programme for Research and Technological Development (FP7 pdf) lasted from 2007 to 2013 and had a total budget of over €50 billion. 64% was allocated to the Cooperation programme for research carried out in ten key thematic areas, such as health, food, energy and transport, and 15% of the budget was allocated to the Ideas programme to support frontier research solely on the basis of scientific excellence… carried out in any area of science or technology, including engineering, socio-economic sciences and the humanities. To implement the Ideas programme, the European Research Council (ERC) was set up. The ERC positioned itself (pdf) to raise the ability for the EU to compete globally, to enable competition between researchers across borders within the EU, and to fund research at the frontier of knowledge production. This definition was intended to overcome the dichotomy between basic and applied research, long considered to be obsolete, especially in the most dynamic and advanced fields of scientific knowledge production. It also intends to make Europe and its research institutions attractive for the best researchers from anywhere in the world and to pursue cutting-edge research with its enormous potential for societal benefit and economic growth. Of the 4352 projects funded, 23% went to the top ten host institutions, which comprised five national laboratories, four UK universities, and one Israeli university.

By May 2020 the ERC had funded 7702 projects in Life Sciences, Physical Science and Engineering. Approximately 70% of these use application related words in their project summaries.

There were two major research and development programmes in the recent European Framework FP7: Cooperation, which invited proposals for collaborative projects with clearly defined scientific and technological objectives and specific expected results, and Ideas under which the ERC, representing the scientific community, invited proposals from the scientific community to be selected on the basis of excellence. The knowledge produced from these projects takes the form of research publications that are available in the literature.

There is considerable overlap between the 7914 Cooperation programme projects and the 4568 Ideas programme projects on CORDIS, the EU research results database. For example, the table at the right shows in graphical form the number of funded projects under each programme that contain references to each of three artefacts: carbon fibre, photovoltaic solar cell, and nanowire. The proportion of projects in the two programmes is different for each artefact, as are the histories of their developments. The table includes plots showing the proportion of research publications which had words related to each artefact in their title, as recorded in the Web of Science database. The years of some relevant technical milestones are also marked. The artefacts originated 30 to 100 years before the periods shown, but research interest was negligible until a technical development opened up new uses, after which research increased dramatically.

So, for carbon fibre the first high strength, high cost fibres were used in specialist aerospace applications after 1965, accompanied by a peak in research publication, which fell off again as the limitations and cost of the material became apparent. The development of lower cost, more reliable production and wider application by 1990 was followed by rising publication rates. The pattern shows technical advances posing new questions, and answers being sought in research. This is in accord with the balance of FP7 projects being towards Cooperation, with its clearly defined objectives.

Research into photovoltaics has had three such developments: in the 1960s with semiconductor technology, in the 1980s with mass production, and after 1990 when the global output capacity began to double every 2.4 years. Some of the research was into questions posed by applications of the usual silicon system; some is into improvement of that system, and some is into alternative systems that might offer lower price, higher efficiency or new opportunities for deployment. These alternatives are speculative, in that there are known and unknown obstacles to making them economically viable, and some are more speculative than others. This is in accord with the more equal balance between Cooperation and Ideas programmes in the proportion of FP7 projects that concern photovoltaics.

The development of nanowires from their precursor in the 1960s was followed by further technical developments in nano tools in the 70s and 80s, and finally the dramatic growth of research after new techniques published in 1999. Unlike the developments mentioned above, the latter breakthrough was made in a university research laboratory, but before his work was published (Morales and Lieber) the lead author had moved to a national laboratory, because, he said I wanted to work with short-term deliverables rather than on projects with results expected somewhere in the far future. Nanowires are not yet in use, but there are at least eight areas where applications in electronics are envisaged. This large volume of speculative research is driven by the hope for ultimate technical outcomes of large value. This is in accord with the dominance of the Ideas programme over Cooperation in the proportion of relevant FP7 projects. It also illustrates the increase in technical involvement of university research.

It is conspicuous in these examples how low the publication rates fall when there is no stimulus from actual or prospective application. These artefacts embodied new materials, new processes within them, and new states of matter, all profoundly interesting to the curious mind. That curiosity did not produce research activity: the inspiration to look deeply into these unknowns came from application. The Frascati Manual characterises basic research as having a longer lead time to application, and a broader potential field of application than applied research. It seems from the above examples that this does not distinguish the research funded under the Ideas programme from the research funded under Cooperation. The programmes can instead be distinguished by how great the risk is that they will ultimately fail to contribute to any application, and by how high the potential returns might be if they succeed.

This distinction is familiar from planning any technical development programme: an essential part of planning is choosing the extent of the research, and how far to go into unfamiliar or unknown areas. Parts of the research are expected to be relatively straightforward, and others more speculative, with longer term, less certain outcomes. The latter parts are described by the ERC as cutting edge. These may fit the Frascati Manual description of Basic Research as being undertaken primarily to acquire new knowledge of the underlying foundation of phenomena, but they are not part of a systematic exploration of those foundations; they are instead part of a framework of technical development.

The artefacts and research topics discussed above all relate to industrial products, but the framework of technical development is not restricted to these. A climate projection, or a health intervention are both artefacts, so a programme with the construction or improvement of one of them as an objective is also a programme of technical development. Such programmes have a wide variety of characteristics and techniques, but all will have some strategy for choosing what to look into, and how far to go into it.

Science and technology studies

Mirowski & SentThe Commercialization of Science and the Response of STS, Philip Mirowski and Esther-Mirjam Sent, in The Handbook of Science and Technology Studies, edited by Edward J. Hackett, and Michael E. Lynch, MIT Press, 2007 Ch.26 identify three American [US] regimes of science organisation in the 20th century. They describe how from 1960 to 2000 the proportion of funding of R&D by business rose from 30% to 70% as the post-war regime of public funding gave way to what they call The Globalized Privatization Regime. Given that applied research in the US receives more than twice as much funding from business as basic research does, this indicates that the fears expressed in EF, that basic research will be driven out by applied, may have been justified. However, the figures from the US National Science Foundation [NSF] indicate that the basic/applied funding ratio actually rose from 62% to 86% between 1970 and 2015. The NSF uses the Frascati Manual definitions, but given Benoit Godin Benoit Godin, Measuring science: is there 'basic research' without statistics? Social Science Information 2003, 42(1), pp. 57-90.'s observation that the use of the term basic research has shifted over the years during which it has been employed in funding decisions, it may be that the figures do not portray what is actually going on in the field. Those seeking funding for research are motivated to identify it as whatever will maximise their chance of success.

Review

The examination of a large contemporary research programme showed how its projects could be characterised as more or less speculative parts of technical developments. This indicates continuity with Rosenberg's observation that in the past, technological knowledge has preceded scientific knowledge, and confirms Jones' suggestion that incremental improvement may still be the major source of innovation, and follows the guidance from other authors that research should not be separated into opposite types such as basic or applied.

This view of research as part of a technical framework contrasts with the ERC's assessment of research proposals in terms of excellence. This excellence is determined by peer review among the scientific community:

The principle of excellence only would remain an empty shell, if it were not implemented in a robust and reliable way. … The crucial element in its peer review system, however, has been the Scientific Council’s responsibility for selecting the panel members.

The outcome of the peer review process depends on the community and its objectives. These objectives clearly include the achievement of technical advances, as well as perhaps the systematic exploration of knowledge. Even within a technical framework, it is still possible that the latter objective guides speculative research in a direction which results ultimately in key innovations. If that were the case, the view expressed on Introduction and Conclusions that science precedes technology would have some substance.

To examine that possibility, more material on science and its community has to be consulted.

Defining science

Back to the four sources

Uses or definitions of the term science from the four sources are in the bottom row of the table. Despite its title, EF does not say what it means by science; the term is used to cover all forms of technical or scientific activities, so EF does not contribute to distinguishing one from the other. Wikipedia defines science as building and organising knowledge, and Britannica as pursuing knowledge of general truths and fundamental laws, as well as being a system of knowledge. The latter definition is problematic because it does not specify its point of reference: a system of knowledge might denote the contents of journals or books, or the knowledge of practitioners, or the knowledge of researchers, or the knowledge taught and assessed by disciplines that identify themselves as sciences. The Frascati Manual avoids these problems of definition, and refers only to research and development, regarding it as the pursuit or building of knowledge. It does not use science as a general term, but instead it refers to scientific research, disciplines, journals, knowledge. Its description of basic research as undertaken primarily to acquire new knowledge of the underlying foundations of phenomena and observable facts and saying that it analyses properties, structures and relationships with a view to formulating and testing hypotheses, theories or laws is similar to Britannica's first description of science.

There are problems also with distinguishing science as being the pursuit of general truths, fundamental laws and underlying foundations, rather than any other sort of knowledge. It implies the pursuit of a goal which does not exist yet and which will take an unknown form. The characterisation of the process as a pursuit is susceptible to wishful construction in retrospect (see for example WigglesworthWigglesworth, V. B., The contribution of pure science to applied biology. Ann. appl. Biol. 42, 34-44 (1955), https://doi.org/10.1111/j.1744-7348.1955.tb02408.x p39). For the pursuit to be guided by existing knowledge, that knowledge has to have a clear structure. This may be identified for limited fields of science at a textbook level, but at a research level it is far from clear. Even in with the benefit of hindsight it is often unclear as to where the new knowledge has come from. These problems have been the subject of some celebrated enquiries.

Philosophical scientists

Polanyi (pdf) proposed that:

The scientific method was devised precisely for the purpose of elucidating the nature of things under more carefully controlled conditions and by more rigorous criteria than are present in the situations created by practical problems. These conditions and criteria can be discovered only by taking a purely scientific interest in the matter, which again can exist only in minds educated in the appreciation of scientific value. … No important discovery can be made in science by anyone who does not believe that science is important—indeed supremely important—in itself.

Kuhn (pdf) stated

If science is the constellation of facts, theories, and methods collected in current texts, then scientists are the men who, successfully or not, have striven to contribute one or another element to that particular constellation. Scientific development becomes the piecemeal process by which these items have been added, singly and in combination, to the ever growing stockpile that constitutes scientific technique and knowledge. …Both normal science and revolutions are, however, community-based activities. To discover and analyze them, one must first unravel the changing community structure of the sciences over time.

The philosophical scientists offer a way forward: for Polanyi, scientific interest can exist only in minds educated in the appreciation of scientific value. He explains scientific value in the following terms: An affirmation will be acceptable as part of science, and will be the more valuable to science, the more it possesses: certainty (accuracy), systematic relevance (profundity), and intrinsic interest. The less exact sciences, such as biology, score lower on the first criterion but higher on the last. The first criterion might exclude arts and humanities, but not technology. Scientific value is routinely assessed by referees consulted by scientific journals to decide on publication (p144), and

…this difference between pure science and technology…is unquestioningly embodied in the general framework of higher education…in the current distinctions between pure and applied chemistry… etc., in the description of university chairs, journals and international congresses; it determines the conditions of employment of scientists in universities on the one hand and industrial laboratories on the other; (p190)

For Kuhn too, the development of science takes place in discourses within communities of educated minds. The discourses have many forms, such as the production of journals and conferences, the publication and reading of papers, attendance at seminars, and informal conversations within structured teams. A threshold of relevance to the discourse has to be met for a finding to be approved for publication and so enter a discourse. Kuhn located these communities in disciplines identified as sciences according to their academic qualifications, societies and journals. He describes a discipline as being established when it has formed a disciplinary matrix (or paradigm) consisting of laws, models, values and exemplars.

Polyani and Kuhn are turning attention away from the intrinsic qualities of activities and towards the social conditions under which they take place. This implies that the distinction between science and technology lies in their being practised in separate disciplines. The separate identity of disciplines as academic enterprises has been challenged, however, both from the examination of historic cases, and by analyses of current developments. According to Lenoir: …scientific and engineering disciplines embed within them the structure of the general power relations in society, while each disciplinary field's own activity of selection and indoctrination contributes to sustaining that structure. In a review of eight diagnoses of the changing state of scientific research since 1997, HesselsLaurens K. Hessels , Harro van Lente, Re-thinking new knowledge production: A literature review and a research agenda, Research Policy 37 (2008) 740–760 reports that all of them are concerned with Choice of research agenda (research content) and Interaction with other societal ‘spheres’ (industry, government). These challenges imply that the identification of research disciplines, and the location of science among them, is not straightforward.

Kuhn recommended the study of academic correspondence and linkages of citations to see how research communities operate at a more detailed level. The study of citations and their linkages has developed extensively since then: The core of scientometrics consists of indicators of scientific output: publications and citations (van BochoveC.A. van Bochove, Economic statistics and scientometrics, Scientometrics 96, 799–818 (2013)). These indicators might provide statistical evidence as to the existence and operation of disciplines.

Scientometrics

The size and scope of research communities are displayed in the Science Citation Index (SCI), which groups 14,492 journals into 178 categories, from mathematics with 314 journals to andrology with seven. There are ten categories with technology in their name, comprising 631 journals, but seven of these categories also have science in their name, leaving only 92 journals in categories that are designated solely as ‘technology’. More than 70% of the SCI journals are in categories where research and practice are combined, such as medicine and engineering. The remainder are in the disciplines of biology, chemistry, physics or maths. Within these, about 17% of the journals are in categories that indicate clearly that they are applied science, but the rest give no indication of how concerned they are with applications. The way that the SCI categorises its publications indicates no clear division between scientific and technical publications.

The Web of Science database of publications looks more promising: it groups 157 research areas under five general headings, one of which is Technology. However, six out of 21 areas under this heading have science in their name, and the general heading of Life Sciences & Biomedicine makes no distinction between medical science and clinical practice. In the ScienceDirect database, Health Science includes clinical practice, and half the subjects listed under Physical Sciences are technical.

It seems that the clear demarcation between science and technology relied on by Polanyi is not apparent in the categorisations of academic journals. These categories might not be representative of how research communities actually operate: there may be groupings of disciplines that can be identified by studying how researchers depend on each others work. An investigation (Leydesdorff pdf) of the citation links between Science Citation Index journals shows that their categories can be clustered into fourteen broad groupings, as shown by the coloring in the figure at left. The lines show the existence and strength of links between the groupings: the connections are more striking than the separations. There is no suggestion of any division between science and technology.

Another analysis looks at how the research communities are formed at a more detailed level, without employing the categories in the databases. Of 58.5 million publications from the Dimensions database, 38.8 million publications had links, forming 2890 clusters. Each cluster was then assigned to one of five main areas of research, as shown by the coloring in the figure at right. The remaining 19.7 million publications were disregarded: most of them had no citation links with other publications, and many of them had no significant scholarly content. These clusters indicate the formation of communities within which a discourse is taking place. Engineering and physical sciences form one main discipline, biomedical and health form another; none offer a division between science and technology.

A study by Robert Tijssen Robert J.W. Tijssen, Discarding the ‘Basic Science/Applied Science’ Dichotomy: A Knowledge Utilization Triangle Classification System of Research Journals, Journal of the American Society for Information Science And Technology, 61(9):1842–1852, 2010 sought to identify high industrial or clinical relevance in a journal by a count of the fraction of industrial or clinical affiliations among its authors. The content of these journals might then be identified as conveying the output of practitioners, as distinct from basic researchers. A small number of journals did have significantly higher fractions, but the median value of the fractions over all journals was only 1⋅2%. So the published body of literature is overwhelmingly dominated by the output of researchers rather than practitioners. Research activities in the business enterprise sector consumed more than 50% of OECD research funding in 2013, so it seems that its output should not be measured by publication in academic journals. This problem of measurement is also found in a study of cross-national research productivity comparisons (Aksnes et al.Dag. W. Aksnes, Gunnar Sivertsen, Thed N. van Leeuwen and Kaja K. Wendt, Measuring the productivity of national R&D systems: Challenges in cross-national comparisons of R&D input and publication output indicators, Science and Public Policy, 44(2), 2017, 246–258. doi: 10.1093/scipol/scw058).

The researchers who are represented in academic journals belong to academic institutions: according to the author affiliations in the Scopus database in 2018 over 85% of publications are associated with authors in the HE sector, and these together with national academies, institutes and research centres account for all but 2% of publications. The scientometric evidence indicates that although science and technology disciplines cannot be distinguished within the network of published research, there is a sharp distinction between published and unpublished research outputs, and the organisations that originate them. Since academic research output is measured by research publication, and almost all such publications are produced by members of academic institutions, we can associate academic research with the development of the network of published research connected by citations.

Although definitions of science (as distinct from technology) have proved unsatisfactory, we have at least identified a strong relationship between research conducted in academic institutions and research that is published. Studies of this published output indicate a differentiation only at the broadest level into research disciplines, and that these disciplines are not identifiable as science or technology. There are too many internal connections, and mutual citations, for any more detailed segregation. The institutional relationship between academia and published research suggests the need to look more closely at what distinguishes academic research from others kinds of research and development.

Academic research

Although academic research is too speculative to attract short term commercial investment, it has many strengths. It is established in universities and national institutes and academies, with staffed and furnished laboratories and information resources. They organise the system of publication and review which defines and validates knowledge, they are highly competitive within and between them, they have high cultural status and are prized in developing countries, and they are the predominant influence on curriculum and assessment in education. This means that the research community has public visibility, and it is the subject of presentations external to the community. Its presentation in education is particularly significant, as it is encountered by entrants to the community: this is examined on the Technology and science education page. The community has a number of operational principles, such as attribution, autonomy and coherence. These feature in its external presentation, although the community in practice does not, of course, conform perfectly to them.

Attribution and autonomy

One operational principle is that credit is attributed to researchers by identifying the authorship of each publication. This convention has been linked to the very origins of science:

The existence of the idea of discovery is a necessary precondition for science,… With the emergence of the idea of discovery and the consequent development both of priority disputes and of the determination to link every discovery to a named discoverer, something that is recognizable as modern science begins to appear for the first time.

(WoottonDavid Wootton, The invention of science : a new history of the scientific revolution, Penguin, 2015 p105). The prospect of recognition for discoveries remains the raison d'être of the academic researcher (Bayvel). Not only discoveries, but also laws, procedures, equipment and even units of measurement are identified with the name of a supposed originator or relevant historical figure.

In principle the publication and review process adds only new, significant findings to the literature. Where the authors have made use of a previous finding the publication containing it is cited, and so the previous finding becomes attributed to the authors of the publication which reported it. Researchers receive social and financial rewards for achieving large numbers of citations. The motivation for receiving this sort of credit takes priority in determining the conduct of research. This is only overridden in unprecedented circumstances: It’s not a time to get your name on an academic paper. It’s a time to find a vaccine. (Dold 2020). The motivation is so strong that it has degraded the quality of academic research output, and may make substantial parts of it unreliable (Alok Jha) or insignificant (RoRI). Considering the strength of this motivation, it is not surprising that academic researchers have claimed that the credit for technical innovation should go to academic research, and have exhibited bias in researching that claim.

Another operational principle of the academic research community is its autonomy. This too is a foundational principle which endures, from the Galileo affair to the debates on Climate change denial. This autonomy protects the community against any demand by outside interests that it suppress or alter its research findings when they relate to matters of practical importance. Autonomy is embodied in research policy: of the Ideas programme the ERC Scientific Council says: In order to set up the evaluation system for such an ambitious undertaking, best practice dictated that scientific strategy should be put into the hands of scientists. Thus, the ERC Scientific Council was entrusted with this task and its autonomy was to be guaranteed by the EC. It appears that the ERC is using this autonomy to protect curiosity driven research at the frontiers of knowledge (HORIZON 2020) against an opposite kind of demand, the demand that it should relate to matters of practical importance.

The principle of autonomy, together with the stated aim to achieve excellence at the frontier or cutting edge implies that the community has internal criteria which can be used to set priorities to direct research, independent of practical interest. This further suggests that academic research is seen as having a coherent framework within which an academic can locate a frontier, and also estimate the distance of a piece of research work from it. The term excellence is heavily relied on: the ERC paper uses it seven times on one page, and 33 times in its 13 pages. This is a view which relates clearly to Kuhn's disciplinary matrix as the framework and Polanyi's systematic relevance as a measure of excellence.

The pervasive rhetoric of excellence has been severely criticised by Samuel Moore et al:

In the final analysis, it turns out that that excellence is not excellent. Used in its current unqualified form it is a pernicious and dangerous rhetoric that undermines the very foundations of good research and scholarship. … We conclude by proposing an alternative rhetoric based on soundness and capacity-building.

Their call for soundness is answered if research projects are not seen as parts of a framework of academic research which are different in how close they are to a frontier, but as parts of a framework of technical development which are different in how speculative they are. In this framework, research output that is unreliable or lacks value is revealed: any research result has to pass critical tests before being used in practice, and the value of it is revealed by the outcome if and when it is used.

It is clear from the analysis of the FP7 projects above that technical developments are tracked closely by the community. Autonomy and curiosity about nature would not produce this outcome: the question arises as to what does. One answer is that technical development provides the equipment for experiments and data gathering and analysis: Large, advanced S&T infrastructure is a key resource and provides a foundation for top quality scientific research (Overview of the Chinese Academy of Sciences). Another answer is provided by the interconnections between science and technology within the community, and another in the connections between academic researchers and external practice. These connections support academic research which serves actual technology and which aspires to enable prospective technical development.

It appears that the strategy of protecting the autonomy of the community is in practice defending its right to direct the more speculative parts of technical development. It is not obvious why this should be presented as research controlled entirely by internal academic criteria, especially since that appearance makes it superficially less attractive to commercial funding. One answer may lie in the principle of attribution, and the overriding motivation of members of the community to benefit from it. If the community is autonomous, then the credit for discoveries can be entirely attributed to the researchers in it. If any of the credit goes to an input from outside, then the academic researchers will receive less.

Disciplines and doctrine

Another operational principle of academic research is the transfer of knowledge. The researcher is trained to seek out sources of relevant knowledge, and recognise its receipt by citing them. Academic research has been institutionally part of higher education since the time of Humboldt, when new universities and grands écoles were founded. Old universities were subsequently compelled to adapt to the new form in which students learned from current research rather than from scholarship. The connection has grown in strength since then, with 85% of research output coming from researchers in higher education institutions, and the advocacy of research led teaching: Excellent teaching goes hand in hand with excellent research (HEA).

The disciplines (and the definition of science) that are so hard to find in the studies of research can be found through the education connection. The UK Higher Education Statistics Agency is in transition between two classifications of subjects of study. The current system is organised in three levels: the top level has 21 subject areas, of which three are in core sciences and maths, and seven are in other technical areas. Each of these areas is subdivided, forming a second level of about 25 and 50 principal subjects respectively, which are further subdivided into over 700 separate subject codes at the lowest level. The current system is becoming inconsistent and hard to extend. The new system has a single level with randomly assigned codes; their mapping between the systems connects the codes in the old system to about 500 codes in the new one.

The current system is still in use in the UK's UCAS subject guide for choosing an undergraduate university course. It appears that at present, at least, choices for the prospective undergraduate are presented according to a clear disciplinary identity, in which science is separated from technology. Entry to academic research is by way of these courses, and researchers carry their identity with them.

The connection with education works both ways: the HE sector runs the courses that lead to professional qualifications, chooses the prior curricula that are followed to gain entry to them, and influences earlier stages of education as well. This is illustrated in a UK AQA presentation on Progression from KS3 to Higher Education: Curriculum pathways in science, which reports that We are working with a number of universities over the next few years to understand what knowledge and skills undergraduates require to be successful; this will help steer our next generations of qualifications to be as effective as they can be for generations to come. In the earlier stages the learning is organised into disciplines with formal curricula, almost all of which have a clear identity as science or technology, at least in name. The curricula are examined in more detail on Technology and science education, along with some alternatives and developments that have been proposed.

It is not surprising to find that the present science curricula embody the principles of academic research in many ways. They are presented as a coherent framework of concepts, attributed to the discoveries of famed academic researchers. The ideal for the most able student to aspire to is to have their name in the textbooks of future students. The curricula generally embody the assumption that new science produced by academic research is the origin of technical advance.

Review

What is left of the view expressed on Introduction and Conclusions, that sees science and its autonomous systematic exploration of knowledge as the origin of key practical innovations? The material on this page relates to recent research and practice, so it is not of relevance to the quotes concerning the early history of technology. The material does show that the simplicity and confidence of the other statements is misleading. The assumptions about the nature of science and technology they are based on are mistaken. The view of science as a separate autonomous activity does not correspond to the reality of research projects and the community that run them. The studies of knowledge flowing from science to practical applications mainly report ill defined influences that might act in the long term.

In contrast, the explorations of research projects and technical advances show a clear pattern, where academic research is responsive to practical motivations, and performs the the more speculative parts of research and development. Academic research can be distinguished from other research, but each of these types of research is a medley of science and technology. The question arises, what might be the practical consequences for research of changing to this "practice oriented" view of it? The immediate answer is that, since it is only a recognition of how research is actually structured, the new view does not promote any wholesale changes in how it should be carried out. There is no implication that decisions about levels of funding for academic research projects, or which of them are funded, should be made in a new way.

There remains the concern for protecting the funding of basic research expressed by Vannevar Bush and his successors to the present day. In a practice oriented view this becomes a requirement that the developments be done thoroughly, with proper regard for the longer term, and that there is an overall coordination of projects so that cross project and cross disciplinary goals can be identified. These are all matters that are addressed by major research programmes, and those seeking funding from them, at present. They are not always the prime consideration, however: sometimes quick and dirty is preferable to exquisite and too late.

There may be an expectation that a practice oriented view would prejudice against those in areas of research that do not make any connection with a useful practical outcome. They would indeed no longer benefit from the supposition that their practical value is incalculable and indefinitely postponed. They could instead make the case that the fundamental questions they address are of great social interest.

The consequences of changing to a practice oriented view would be significant in education. The prevailing view in the curricula and their design in some countries is that science precedes technology. This is a large and complex issue, and is examined on the Technology and science education page.

Other views on Introduction and Conclusions and Technology and science education turn to accounts of earlier history for support. Many such accounts do indeed share these views, but there is an ongoing debate about them. This debate can only be joined by a detailed examination of the accounts and the evidence. A problem in the history of science and technology does this for some accounts of the early development of steam power.

Conclusions

are drawn together on the Introduction and conclusion page.