The historical synthesis of machine, organism and language

AI’s position in the history of science and technology can be understood as the confluence and synthesis of the three paradigms of machine, organism and language. From this perspective, an AI system has to be understood as a machine that reorganises its own parameters like an organism adapting to its environment in order to model the language of social artefacts. This movement of self-organisation is not simply a technical feature: it reflects the social genesis of technology in general, that is an emergent process in which machines internalise the relational structures of collective practices—structures that include human cooperation, the division of labour across society, and the collective dynamics of language and culture (one should also define language, by the way, as a process of self-organisation). Cybernetics had already framed self-organisation as the capacity of a system to internalise and reproduce the relational structures present in its environment. AI follows this logic at scale: today it takes the form of a vast parametrised machine that maps the dense networks of relations sedimented in data and cultural artefacts. Texts, images and sounds are not isolated entities, but matrices generated by collective practices; semiotically, they are diagrams of social relations. As a matter of fact, what AI automates is not biological intelligence, but the relational structures historically produced by human cooperation (Pasquinelli 2023).

These insights were familiar to cyberneticians such as Karl Deutsch, who in 1951 authored the programmatic essay ‘Mechanism, Organism, and Society’, as an attempt to trace analogous models in natural and social sciences. Deutsch framed the history of thought as the evolution of mechanicism, organicism and historicism culminating in cybernetics. He believed that cybernetics helped to recognise profound ‘analogies […] between communication channels or control processes in machines, nerve systems, and human societies’ and he identified them in the principle of ‘self-modifying communications network or learning net’ (Deutsch 1951: 240). Deutsch’s description of both the project of cybernetics and the ontology of society as a self-organising learning network reads as distinctly technocratic, yet it is quite prophetic of the current predicament of AI. In a way similar to the labour theory of automation and technological development seen above, Deutsch observed that the increasing division and mechanisation of labour had prepared the way for the division and mechanisation of mental labour in cybernetics (and AI):Just as the division of manual labor between different pairs of human hands preceded the division of labor between human hands and power-driven mechanisms, so in a sense the increasing division of intellectual labor between different human minds has preceded today's growing divisions of labor between human minds and an ever-growing array of electronic or other communications, calculating, and control equipment. (Deutsch 1951: 239)

What cybernetics depicted as a self-organising learning network is a machination of the same relational structures that constitute linguistic cooperation and social life. Seen from this angle, the machine, organism and language genealogy acquires a sharper meaning and points once again to the social synthesis that this essay attempts to illuminate: AI is a model of the social manifold.