Preliminary reflection: the other of the instrumental reason? (no answer)

June 25, 2026

We reposition the question of the difference between artificial and human intelligence, and what's special about the latter, as, what is the content of a non-instrumental reason?

Even if it looks like the world started in 2022 with the public release of ChatGPT, it would be precipitate to want to inaugurate a new historical era based solely on that event, revolutionary though it may or may not be. It should serve as an example when, during a different Revolution, the French National Assembly decided to enact the calendrier républicain, marking 1789 as the beginning of a new world history. This lasted for 12 months, after which, because of comings and goings of the political tide, History severely replied by restoring the Nazarene on Saturn's Throne.

If the latest potentially revolutionary technology, artificial intelligence, indeed brings forth a myriad of unresolved technical and philosophical questions, that we may have occasion to explore before the time when human brains are made obsolete in the order of the world, it is fundamental, when gazing at that artifact, the Large Language Model, to take a step back in order to avoid getting lost, dumbfounded in the face of its marvellous chthonic spectacle, and see more clearly and rationally its place in the history of ideas. Mechanistically circumscripted, the artifact of soft and hard matter, metal, electrons, and characters, is the latest spawn in the long devenir of intelligent automata, in the category of which are represented the computer, the calculator, and, archetypically, the clock.

It was René Descartes who, in Les Passions de l'âme, if not first, best developed a method for understanding the operations of body and soul in terms of the movement of parts, fluids, vapours; in a word, a mechanistic system. This system constitutes a constructive paradigm for all subsequent inquiries that approximate and correlate the operations of human and machine in a way that surpasses the physical dimension and invades the spiritual: not because the human is capable of imbuing the machine with his Promethean fire, but rather – this here the ingenious cartesian inversion – because he is himself a machine.

The mechanistic system is a perfectly solid scaffolding on which to build explanations for the physical, chemical and biological movements that allow artifacts to express abstract representations, or thoughts: the hands of the clock answer "what time is it?"; the calculator registers parts, and tells what is the sum. Technological progress appears as a cumulative process in which new features are added to an abstract machine, which the Moderns called an automaton, and we call a computer. Intelligence, in this theoretical paradigm, is built by adding more and more of these mechanisms of abstract representation, or ascribing additional mental qualities to machines, in the language of John McCarthy. His thermostat was capable of believing and deciding, and our LLM is capable of learning and teaching; but without exiting the epistemological frame of mechanicism, we are still in the company of Descartes.

What is more, this abstract machine is theoretically infinitely expansible, by adding more and more features, as long as there is fuel in nature, and demand for its gift in society; there is no reason to suppose the one-to-one blueprint reproduction of a human machine, in all its bits and pieces, to be outside of the realm of possibility. None of this is new to those who lived through the human cloning collective psychosis of the 2000s. If we pose the question in this manner, as, "what does the (abstract) machine lack that is given in the human?" the answer is, "potentially, nothing". Intelligence appearing as a collection of features makes artificial intelligence, in the mechanistic sense, a matter of giving the machine eyes to see, ears to hear, virtually infinite fingers to count, and a powerful algorithm to process the stream of phenomena. But here, in the crevices of the solid frame of mechanistic rationalism, we find the bud of a contradiction that may take root: what is not being questioned here is the effective content of reason itself, as it is put forth by the machine. In the question of intelligence, artificial or otherwise, the gaze should be repositioned, as the object is not the machine, but thought. The question becomes more interesting. Instead of asking what abstract representations can a human express that a machine cannot – not much, it seems now, and the limit gets closer and closer; we can put the question as, what is the determinate content of reason that the machine can operate, the so-called instrumental reason (instrumentelle Vernunft). That should ferry us to the more important question: in the concept of Reason, what are the contents of its other, a non-instrumental reason.

Further Reading