Thinking with Machines

Abstract

The article presents an analytical review of David W. Bates’s book Artificial History of Natural Intelligence: Thinking with Machines from Descartes to the Digital Age (2024), which examines the historical conditions for the formation of the modern concept of intelligence. The review focuses on an analysis of Bates’s research strategy, which makes it possible to trace how the opposition between natural and artificial intelligence took shape within the long history of machine, physiological, cybernetic, and technical models of thought. Methodologically, the review draws on a problem-conceptual analysis of the monograph, a historical-philosophical reconstruction of its argumentation, and a comparison with studies in cybernetics, philosophy of technology, distributed cognition, and predictive processing. It is shown that Bates rejects a linear history of artificial intelligence and instead constructs a trajectory of problems: from Cartesian physiology, early modern theories of the automaton, and the Kantian problematic of judgment, to Babbage, Lovelace, psychophysics, thermodynamics, philosophical anthropology, cybernetics, Turing, theories of technical exteriorization, and models of machine learning. The main result of the review consists in identifying the epistemological significance of Bates’s thesis that the opposition between natural and artificial intelligence is historically derivative. The machine here appears not as an instrument of a ready-made subject, but as an epistemic figure through which the mind made its own operations, limitations, and forms of autonomy observable. In conclusion, it is argued that Bates's book is productive for the philosophy of science, insofar as it allows cognition to be examined as a history of material, bodily, and technical models of thought. At the same time, problematic aspects of the work are noted: the selectivity of the historical composition, the insufficient clarification of the difference between human normative autonomy and machine adaptability, as well as the limitations of the Euro-American framework of the study.