Free Agents (I): the evolution of agency
Tracking the evolution of agency from the origin of life, through the development of nervous systems, and the eventual emergence of free will
As we saw in the comment on Robert Sapolsky's last book, the controversy between those who believe that humans enjoy free will and those who disagree with that notion remains unresolved. The latest contribution to that debate that I have read is Kevin J. Mitchell's latest book, Free Agents (2024).
Let us begin by presenting an admittedly crude, but useful (because crude) classification of the main ideas regarding determinism and free will. On the one hand, there is dualism, a conception of the human being in virtue of which we are endowed with a material entity –the biological organism– and a spiritual one –the soul or equivalent entity– which ultimately governs our thoughts and behavior. Dualism is, logically, intractable from the scientific point of view, since science cannot deal with entities without a physical substratum.
On the other hand, we have determinism, a position which, in its "hard" version, holds that the evolution of the entire universe and of everything that happens in it began from the very beginning of the universe, so that a sequence of innumerable causal relationships has been concatenated by virtue of which each state at a given instant absolutely determines the state at successive instants. Consequently, everything, absolutely everything that happens, is conditioned by the previous states.
It would seem logical that all those who hold this view would deny free will. However, this is not so; the so-called ‘compatibilists’ argue that some degree of free will is possible in a deterministic universe. I confess that I do not quite understand this position.
And there are, thirdly (or fourthly, if we consider compatibilists as a separate category from the determinists as a whole) those who, from a materialistic view of reality, deny that the biological agents –and, in particular, humans– act in a predetermined way.
Mitchell is one of the latter. His approach starts from the analysis of the evolution of living beings and, in particular, of their condition as agents. Agency, in fact, becomes the main notion of his book and only in the second part does he address, in order to deny its validity, determinism, and argues in favor of the existence of free will.
The book traces the evolution of agency from the origin of life itself, through the emergence and development of nervous systems, the subsequent elaboration of decision-making and action selection systems, and the eventual emergence of the kind of conscious cognitive control in humans that Mitchell refers to as free will.
The author points out that we are configured in a way that reflects our history, but that does not mean that we are "hardwired" in the sense of functioning according to a rigid system. We have the capacity for introspection and metacognition, and we can observe our own programming, treating goals and beliefs and desires as cognitive objects that can be recognized and manipulated.
This is not to say that cognitive systems do not have a physical substrate; they do, but their functioning cannot be reduced to that level. We are not a collection of mere mechanisms, because the nervous system works with meanings.
According to the author, the basic laws of physics that deal only with energy, matter and fundamental forces cannot explain what life is or its defining property: living organisms do things; they do them for certain reasons, as causal agents in their own right. They are driven not by energy but by information. And the meaning of that information is embodied in the structure of the system itself, based on its history.
The most primitive organisms give direct, stereotyped responses to every relevant stimulus that is received. The information contained in that stimulus acquires meaning by virtue of its implications for the survival and continuity of the individual or its lineage.
The organism's actions are conditioned –constrained– or, as Mitchell prefers, informed by prior experience. That is precisely the property that differentiates life from other kinds of matter: living things literally incorporate their history into their own physical structure to inform future actions. He argues that, rather than suppressing the freedom of the self to decide at any moment, it is this very process that allows the self to exist. There is no self at any given moment: the self is defined by persistence over time.
This incorporation of the organism's history into its own physical structure is the consequence of the action of natural selection. This factor is of crucial importance because it is the one that provides living beings with ‘a purpose’, that of surviving and leaving offspring.
Natural selection, acting blindly, produces fitness, the adaptation of the organism to the niche. The activities of the system are adjusted in such a way that it manages to endure. This is the object of these activities which, being directed towards the achievement of a goal, are, in fact, functions; that is, functions with respect to the objective of persistence. The existence of that goal imbues living things with properties that matter did not have before: function, meaning and value.
The raison d'être of an organism's adaptations is the organism itself. Since there is thus a goal (the survival and continuity of the genes that carry the information encoded by the organism itself), the objects of the universe acquire meaning and value. The meaning lies not only in the signals that are received, but in their interpretation by the receiver. And that interpretation depends, in turn, on the information –from the Latin in-forma (in the form)– contained in its structure and memory, that is, on its evolutionary history and previous experience.
On the other hand, things are not, per se, good or bad. Those that raise the probability of enduring are good; bad those that reduce that probability. Therefore, value is inherent to the whole system, the one that forms the organism with its environment. It depends on the interaction between the two and the consequences of this interaction over time.
Meaning and value are the currencies that activate the decision-making mechanisms and the selection of actions that emerged as life continued to evolve. Thus, agency is conditioned by the meaning and value of objects; it exists by virtue of these two properties.
With multicellularity came the division of biological labor and thus the need to coordinate internal activities. Neurons proved to be a good system for linking sensory information to action. Once nervous systems emerged, the evolution of a hierarchical architecture made possible a greater degree of interaction of internal and external signals to guide behavior. And this allowed the colonization of more variable and complex environments.
Finally, the emergence of associative learning and long-term memory allowed individuals to transcend their pre-wired instincts and make decisions based on their own reasons. Mammals, primates and, eventually, humans, developed more and more layers of insulation from environmental demands and more and more autonomy as full-fledged causal agents.
In a holistic sense, it is not neural circuits that make decisions, it is the whole organism that does so. It is not a machine computing inputs to generate outputs. It is an integrated self deciding what to do, based on its own reasons. Those reasons are derived from the meaning of the various types of information available to the organism, which is based on its past experience and used to imagine possible futures.
Title: Free Agents–How Evolution Gave Us Free Will
Author: Kevin J. Mitchell
Ed. by Priceton University Press, 2023.
Very commendable Mitchell's effort to argue for the existence of free will, something I also strongly desire. Unfortunately, I am more convinced by Sapolsky's arguments. I find him like a relentless Spinoza with up-to-date science. My hope is that we will make progress in understanding what life (and therefore self and consciousness) is. These times in which every day they try to sell us that both intelligence and artificial life are already here, and we still do not know what one or the other is, it would be good that serious arguments like those of Sapolsky and Mitchell introduce sanity in our thinking.