David Noble: Forces of Production (1984)

“For the process of technological development is essentially social, and thus there is always a large measure of indeterminacy, of freedom, within it. Beyond the very real constraints of energy and matter exists a realm in which human thoughts and actions remain decisive. Therefore, technology does not necessitate. It merely consists of an evolving range of possibilities from which people choose. A social history of technology that explores beneath the appearance of necessity to illuminate these possibilities which technology embodies, reveals as well the contours of the society that realizes or denies them.” Forces of Production is a Marxist analysis of the development of industrial automation. Noble focuses primarily on the machine tool industry, which, despite being relatively small, is essential for the metalworking that is at the core of modern industrial production. The machine tool industry was partially automated through “numeric-control” technology (N/C) that allowed general purpose machinery to be programmed to carry out different actions. It should be noted that Noble’s history of industrial automation is not a history of mainframe computing. Ever since the hysteria over it in the 1950s and 1960s, “automation” has functioned as an unfortunate term whose vague generality covers over the substantial difference between the use of mainframe computers for information processing and the use of cybernetics to control production. As Noble admits, in the early postwar era, mainframe computing was fairly quickly adopted by businesses for data processing, whereas numeric-control, which was developed by different institutions, was only slowly integrated into industrial production. Resisting all forms of technological determinism, Noble begins his book with two chapters of social context. First, he discusses the beginning of the Cold War and the alignment of scientific, military, and corporate interests. He contrasts Senator Harley Kilgore’s failed effort to establish a system that would subject (publically-funded) scientific research to democratic oversight with Vannevar Bush’s more successful promotion of science’s self-sovereignty, the latter paradoxically secured through science’s close relation with military projects and corporate research and development during the Cold War. Noble then turns to “the war at home,” documenting the wave of labor conflicts that appeared in 1945 and 1946 as unions took advantage of the increased numbers and political clout they had gained during wartime production as well as responded to the dangerous working conditions that had been grudgingly accepted during the war (Noble points out that over 88,000 workers were killed in industrial accidents during the war). The Taft-Hartley Act of 1947 substantially diminished the power of unions and labor became progressively weakened as the postwar era continued, but management still actively sought out new ways to control, or even eliminate, labor. Automation of course would come to management’s rescue, playing directly into management’s belief that productivity and profitability could best be improved by granting management even more power over labor and by substituting capital (investment) for labor. Industrial automation was first successfully implemented in continuous-process industries, such as the chemical industries, whose complex production processes both called for and were compatible with basic forms of cybernetic control. The machine tool industries, however, traditionally had involved labor-intensive, small-batch production, a kind of production process that was much harder to automate. According to Noble, early research in the automation of the machine tool industries generated two options. The first was record-playback (R/C), which in theory was not that different from the early gramophone players that allowed users to both record and play back sounds. Record-playback recorded the movement of the machine as it was used by the machinist, and this recording could then be used to replicate the original movement. Record-playback was relatively easy to develop and use, but because it continued to rely on the skills of machinists and did not grant management complete control over the production process, it was ultimately abandoned. Noble points out that Kurt Vonnegut based his novel Player Piano on his experience as a publicist for GE, which attempted to develop record-playback in the early 1950s. Noble argues that Vonnegut’s novel, in addition to not referring to the form of automation that was eventually implemented, is incorrect in imagining that record-playback would ultimately lead to complete obsolescence of the worker. According to Noble, record-playback aimed at augmenting, not completely supplanting, the skills of the worker, and that was its fatal flaw in management’s eyes. The form of automation that was chosen over record-playback was numeric-control, which required the abstract coding and programming of the machine’s instructions. Numeric-control did allow a greater level of precision, though Noble is quick to note that such precision was only necessary and affordable for a small range of military-funded projects. In fact, numeric-control involved such complex programming that for a long period of time it proved too expensive, unreliable, and difficult to install in actual factory conditions. But the convergence of the military’s pursuit of total command and control, science’s preference for abstract models and new methods, and management’s belief that greater management control would necessarily produce greater profits, favored the development of numeric-control instead of record-playback. The Air Force in particular - through funding and the technical requirements it set for its contractors - helped sustain numeric-control for the long period during which its cost was unjustifiable and its complexity almost unmanageable. Needless to say, this kind of Marxist history of technology is out of fashion right now in science and technology studies. Noble's deep research closely follows the historical and institutional contingencies that influenced the development of automation, but his hierarchical explanatory framework, in which the economy is determinant in the last instance, clearly does not allow him to address the full material complexity of his technological subject as thoroughly as might the flattened ontology of the actor-network theory deployed in Atsushi Akera’s similar Calculating the Natural World. But I remain far more sympathetic to Noble’s approach than to Akera’s, which in attempting to maximize the number of human and non-human “actors” included in the web of its historical narrative tends to completely lose sight of the fact that in our society most "translations" between actors are in actuality transactions that are guided by a rather exceptional actor: capital.