I am watching a dog chasing a ball. I see so clearly that matter is not insensate. It is able to yearn, to pursue with frenzy what it craves. Perhaps all matter yearns to become what it can next become, and the evolution of the world is brought about by its ongoing desire. When we say particles are attracted to or repelled by one another, perhaps we should take these terms seriously. Perhaps matter channels itself into different capacities and intensities to desire. How else to make sense of the clear fact that once there was a zygote, before that carbon and hydrogen, and now a frantic chase?
People are eagerly discussing whether ChatGPT is intelligent, and if not, they are making predictions on when it might be. ChatGPT’s performance is not arising out of the attractions and repulsions of its matter, but by a structure imposed on it. By contrast, whatever problem solving abilities dogs have, they are in service of what dogs desire. In life, intelligence is how organisms find ways of satisfying what they yearn. It is completely interconnected with and dependent upon an emotive tug. Without this tug, there is no incentive. The animal sits listless, –almost like a computer. Granted, if all matter yearns, then the electrons and silicone and what not inside a computer will be compelled towards certain ends. But these ends are being funnelled by the organisation of the hardware and the constraints of the software. Perhaps the electrons desire to move down the wires. But the disconnect between the apparent function of the computer at the level we interface with it, and its material process is most evident when the computer breaks down owing to a ‘malfunction’ of some inner component. It no longer does the apparently intelligent things we want it to, but it may in fact be an instance where matter is achieving change through its own quietly persisting willing and achieving. One might suggest that it is no different from a cell ‘breaking down’ and becoming cancerous, but the analogy is flawed. The activity of cells produces the multicellular organisms that they in turn depend upon. A cell malfunctions when it no longer reliably brings forth what it has co-created. The material in a computer is not disrupting its own creation when it burns out a circuit.
In the tradition of enactivism, biologists sometimes suggest that cognition and living are the same process (ex Maturana and Varela). Cognition makes the self, and self-making is cognitive. The intuition when seeing the powerful impulsive obsession of a dog and contrasting it with the passivity of ChatGPT opens for me thoughts that such self-making (or ‘autopoiesis’) is driven by desire. Schopenhauer spoke of a ‘will’ in nature, and Bergson of a desire or volition that creatively compels evolutionary process. And yet, the idea seems odd to modern ears (even New Materialists feel a need to de-phenomenologise desire), as science has decided (without evidence one way or another — what evidence could settle the question?), that matter cannot yearn unless it happens to be a human, or perhaps a mammal. But how molecules, operating according to physicochemical ‘laws’ evolve the ability to thirst for anything, is quite impenetrable. Wherever we look in physiology, we see mechanisms of causal interaction between parts. We even see the mechanisms underlying such thirst. But this does not explain its urgency, intensity, feeling and power.
I should also say we don’t even know where to begin in thinking about how to create a computer programme that desire. In science fiction depictions, it often just kind of happens. The problem may be that unles matter itself is organising into hardware and software, we will be stuck simulating desire through imposing constraints and rules. And so, the desire in ChatGPT remains extrinsic, in the minds of the people creating it.
One might suppose all matter is affective. While we say that particles yearn or have desire, this is not meant to suggest it is at all like the spatiotemporally complex and textured kinds of affects that multicellular organisms can have. The difference between the kind of desire a particle has compared to a human may be as wide as the structural and organisational difference between them. It is simply to suggest that, just as a human and a particle are both material, albeit vastly contrasting in scale and complexity, so too are they both feeling, and with just as much difference.
It is not scientific to assume that matter is feeling or unfeeling because we have no empirical evidence either way. And yet, this is nevertheless a topic that matters for science education. Why? Because we do have theories and hypotheses about the evolution and development of feeling (since Darwin many biologists have written about it), and also because it is striking that all our scientific progress remains inconclusive on this point. It tells us something about science, and about nature, when empirical investigation cannot make concrete progress in certain realms of inquiry. The question unfathomable marks science opens up are just as important ‘products’ as the predictive theories it generates. Particles may well be completely insensate, but then it is a powerful mystery how it is possible that they organised themselves into feeling.
But the most important reason it matters, I think, is twofold. On the one hand, the misidentification of AI with intelligence is rooted in a sharp dualism between cognition and a feeling body, and each time we succumb to this separation in thinking about entities in the world, we further detach our thinking minds from our own feeling bodies. We risk becoming more like a computer at a time when we need to pay attention to what our own hearts yearn for, and reorienting thinking in such service. On the other hand, such misidentification also continues to drive a wedge between humans (and computers) and the rest of the biological world. The latter may not have the computation complexity of a computer or a person, but calling AI intelligent and not a tree or bird does much damage to our relationship with other species.