A Brief Alternative History
Chapter 1: Prehistory
The discourse on technological innovation did not emerge from nothing. It emerged from, and as a reaction to, the previous discourses and concepts on which it is constituted. One such concept is technology (a good that embodies knowledge). Technological innovation is the use of technology in practice, not the technology itself – although it is often used in that sense. Second, technology gave rise to a plethora of concepts with the adjective “technological” in their titles. Before technological innovation came on the scene, several other concepts or terms paved the way for a broader understanding of technology, namely the consideration of its effects on society: technological unemployment, technological change and technological progress.
At the heart of the technology question at the beginning of the twentieth century was labor issues. The first term coined with technology in its name, or at the least debated extensively, was technological unemployment. The term emerged in the context of the depressions of the 1920s. Following the crash of the 1920s, “the term ‘technological unemployment’, first coined by economist Sumner Slichter in the second half of the 1920s (Slichter, 1928), became current”, as economist Harry Jerome of the National Bureau of Economic Research (NBER) stated in Mechanization in Industry (1934). “Then, with the close of the decade, a prolonged depression set in … There came a sharp emphasis upon, if not an exaggeration of, the actual and potential adverse effects of the machine” (Jerome, 1934: 4–5). Technological unemployment is unemployment due to the introduction of “machines” or “inventions” in industry – two key terms of the nineteenth century. Until then, at least to classical economists, unemployment due to machines or inventions was reabsorbed automatically, so it was said (see Gourvitch, 1940). Then, in the early 1930s, the economists added “time” to the equation. The question was then whether technological unemployment is permanent or temporary.
The debate ended with a relative but dominant consensus. Technological unemployment is temporary. It is a matter of job displacements. A job lost somewhere is gained elsewhere. “For every man laid off a new job has been created somewhere”, claimed Paul Douglas, a major theorist of the time (Douglas, 1930: 930). Not all inventions cause unemployment. Some create totally new industries and new types of employment. Technological unemployment depends on types of invention (labor-saving or capital-saving) and types of job affected (skilled, semi-skilled, unskilled).
By the 1950s, few doubted that inventions are beneficial to the economy:
Not a single witness [as the congressional Subcommittee on Economic Stabilization recounted in a report on automation and technological change] raised a voice in opposition to automation and advancing technology. This was true of the representatives of organized labor as well as of those who spoke from the side of management. None of the evidence available before the subcommittee supports a charge that organized labor opposes or resists dynamic progress. (US Congress, 1955: 4–5)
I did not once hear during the session the words “technological unemployment” [claimed Gerhard Colm, Chief Economist at the National Planning Association, to the participants at a National Science Foundation conference on research and development and its impact on the economy]. It is very remarkable that in the midst of a serious recession, with a large number of unemployed, no one suggested that there is any relationship between technological advance and unemployment. For the present period, I do not think this is an omission … Research is adding to demand and thereby creating more jobs than it is displacing. (Colm, 1958: 152)
The term technological change emerged out of the debate on technological unemployment. The term was coined in the 1920s (Alford, 1929; Hansen, 1921; Kuznets, 1929, 1930). It entered the vocabulary widely in the late 1930s. The National Research Project (NRP) on “Reemployment Opportunities and Recent Changes in Industrial Techniques” of the Works Progress Administration (WPA) is a key factor here. Between 1937 and 1940, the NRP conducted surveys of workers’ experiences with machines in industry, with over 1,000 assistants in the field (Weintraub and Kaplan, 1938). Under the direction of David Weintraub, a young economist from the National Bureau of Economic Research (NBER), the NRP produced dozens of studies in collaboration with the NBER and several government departments and agencies (Department of Agriculture, Bureau of Labor, Bureau of Mines). To the NRP, technological change is change in industrial techniques of production, and it is measured like technological unemployment is measured: output per man-hour (Weintraub, 1932). Changes in productivity indicate changes in industrial techniques.
To the NRP, technological change is a source of positive effects on society (outputs, productivity, economic progress), not just negative ones (unemployment). But in order to realize the full benefits of technological change on the economy, there is a need for more investment in capital (industrial techniques). Here, the NRP was espousing economist Simon Kuznets’s retardation thesis and Alvin Hansen’s secular stagnation theory: investments in capital were more numerous and revolutionary in the past than they were at that time, hence the need for more investments (Hansen, 1939; Kuznets, 1929, 1930). This is a matter of business cycles. Joseph Schumpeter explained periods of recession and periods of prosperity based on the spread of such revolutionary innovations (Schumpeter, 1939).
From that time on, technological change became a catchword whose meaning includes (or oscillates between) change in industrial techniques (the technical) and the changes to society and the economy (the change) due to industrial techniques. The key use of the term is the latter. The study of change in industrial techniques per se is left to historians (of invention).
In the 1940s technological change acquired a more restricted meaning that became very influential among economists. Defined by way of what is called the production function (combining input – labor and capital – to produce output) (Brozen, 1942; Schumpeter, 1939), the study of technological change became an industry. Econometric studies were published by the dozen in the economic literature. Certainly, other representations of technological change existed. To sociologists, anthropologists and philosophers, technological change is broader than just the use of industrial techniques in production. It includes any invention to which people have to adjust: farm practices, transportation, communication, and so on. But these disciplines did not develop a discourse on technological change per se. For example, the US Social Science Research Council, supported by sociologist of invention William Ogburn, tried to get into the field in the late 1940s, but it failed in its efforts (Godin, 2019). By that time, technological change had an economic (and mathematical) connotation that made it hard for sociologists to follow.
By the 1950s–60s, technological change got on to the public agenda and became “a modern sounding term”, as one of the chapters of the US Commission on National Goals put it (Watson, 1960: 193). Organizations like the OECD, UNESCO and the International Labour Office, as well as policymakers and advisers, espoused the concept (together with that of automation) and developed thoughts for policies on technological change.1 By that time, technological unemployment was no longer a fundamental issue. “Technological advance is a key element in economic progress” (“increases in our standard of living”), as the US Council of Economic Advisers put it, “Our objective should be to foster and encourage it” (US Council of Economic Advisers, 1964: 103, 85).
Practitioners and scholars also began to suggest that a key factor or condition of technological change – namely, science – has not been considered enough. Science is the source of technological progress. The concept of technological progress is not easy to define. The meaning oscillates between progress in technology (industrial techniques) and economic progress due to technology (growth in output) (Fabricant, 1965: 5–6; Kennedy and Thirlwall, 1972: 12). The two are often considered to be synonymous. “Economic progress which results from a change in knowledge is known as technological progress” claims Vernon Ruttan, economist at the University of Minnesota and a prolific scholar on technological change in the 1950s and after (Ruttan, 1954: 1).
The first to espouse a discourse on science and progress in the twentieth century were US industrialists such as Frank Jewett and John Carty in the 1920s, supported by the US National Research Council (Godin, 2009; Godin and Schauz, 2016). The Department of Scientific and Industrial Research (DSIR) in Britain held a similar discourse (British Department of Scientific and Industrial Research, 1927). Subsequently, a whole literature developed on the management of industrial research. The discourse culminated in Vannevar Bush’s Science: The Endless Frontier, which proposed the creation of a National Research Foundation to support basic research publicly (Bush, 1945):
Advances in science … when put to practical use … insure our health, prosperity, and security as a nation in the modern world. (p. 11)
Basic research … creates the fund from which the practical applications of knowledge must be drawn. New products and new processes do not appear full-grown. They are founded on new principles and new conceptions, which in turn are painstakingly developed by research in the purest realm of science … Basic research is the pacemaker of technological progress. (p. 19)
To make sense of the discourse, industrialists developed models, as we call them today, picturing technological progress in terms of a sequence called “diagram” (Mees, 1920), or a “research cycle” (Holland, 1928), or “flow chart” (Bichowsky, 1942; Furnas, 1948), or simply “stages” (Stevens, 1941) (see Figure 1.1). These sequences culminated in that of economic historian Rupert Maclaurin. In the 1940s, Maclaurin from MIT, developed the first-ever scholarly research program on technological change – over the years, Maclaurin shifted to the term “technological innovation”. Maclaurin gave central place to research as a factor of technological change, explaining technological change as a “sequence” from research to use: pure science, invention, innovation, finance, acceptance (Maclaurin, 1949, 1953). By the 1960s, some called this sequence the rational or “common-sense view” (Schon, 1967: 54, 140). Subsequently, the sequence came to be called the “linear model of innovation” (Langrish et al., 1972; Price and Bass, 1969).
In the late 1950s and early 1960s, reviews (see, for example, Nelson, 1959a), conferences (see US National Bureau of Economic Research, 1962) and textbooks on the economics of science (see Bright, 1964) began to appear. In the end, most scholars were concerned with research (as source of or input to economic progress), not innovation (as output of research activities, which is taken for granted). “This [review] paper”, claimed Richard Nelson of RAND Corporation, “is not concerned with innovation. It is concerned only with how inventions occur” (Nelson, 1959a; see also National Bureau of Economic Research, 1962: 4). Nelson discusses but criticizes and minimizes innovation issues (the “overemphasis on the market or ‘demand’ side” of invention; social needs explaining invention) espoused by sociologists like William Ogburn and Colum Gilfillan, and economists like Jacob Schmookler, at the cost of de-emphasizing research. What motivated Nelson to make such a move? The belief in basic research as a prime mover of economic growth. In another paper in the same year, Nelson states: “It is basic research, not applied research, from which significant advances have usually resulted” (Nelson, 1959b: 301). “It has become almost trite to argue that we are not spending as much on basic research as we should. But, though dollar figures have been suggested, they have not been based on economic analysis … How much should we spend on basic research? Replacing the X of the familiar literature on welfare economics with ‘basic research’ provides the theoretical answer” (Nelson, 1959b: 297). Nelson contrasts the private costs of research activities to the “obvious” social returns of research to make a case for more basic research – social returns became a much-studied topic later on (Griliches, 1958; Mansfield et al., 1977; Minasian, 1969).
Figure 1.1: Clifford Furnas’s flow diagram from research to sales