The universality of the machine: labour process theory and the absorption of the skills and knowledge of labour into capital

This article contends that deskilling is best understood not as a distinct phenomenon, but as a component of a process Marx (1993: 694) called ‘absorption’. Absorption involves not only the extraction of capacities from labour but also their implementation in machines. The article reads Braverman’s (1998) analysis of Taylorism as a demonstration of how absorption entails a specific labour process of its own, which I call the absorption process. The nature of the absorption process is contingent on many social factors. This article focuses on a technical factor: the particular machines used to implement captured skills and knowledge, called here the infrastructure of absorption. Since technological capacities are ever-evolving under capital due to the continual revolutionizing of the means of production, infrastructures of absorption change over time and this necessitates new absorption processes. Braverman (1998: 132) pointed to a qualitative change in absorption with the digital computer, which he described in terms of a new ‘universality of the machine’. While Braverman rightly pointed out the computer as a novel infrastructure, he did not discern qualitative changes to the absorption process, seeing instead the extension of Taylorist processes of capture of knowledge and skill. I contend that a qualitative shift has become apparent since the rise of machine learning in around 2015. Machine learning enables a different absorption process of emergence which does not require the codification of captured knowledge. Much labour process theory (LPT) (and adjacent) research presumes that deskilling and automation operate in terms of a process of capture, however, I show that emergence presents qualitatively different means for both. I suggest that the infrastructure of machine learning presents the possibility of task-agnostic automation .