Philosophy and Artificial Intelligence #3 — Freedom of Will
See the previous posts here & here.
What does the freedom of will mean?
The entire concept of free will means that an agent making a choice 1) could do otherwise and 2) is aware of making the choice. Therefore they are responsible for what happened as a result of their choices, because without them the timeline would look differently (better or worse). If we were aware of making choices and not being able to do otherwise, or not being aware of them, the concept of personal responsibility would not make any sense.
But what does it mean exactly that the agent could “do otherwise”? In a given time T1 there is a state S1. If the agent that acts in T1 is free in their actions, that means we have a set of possible states {S2,S3,S4,…Sn} and the fact that only S2 was observed in T2 means it was caused only by free action (or by the lack of action) of an agent.
But how do we know that states {S3,S4,…Sn} were possible? Since the S2 is the only one that was observed, maybe it is the only one that was actually possible? If you observe an ant going somewhere it also fits the description above, and we do not attribute it with “freedom of will” — there is only the probability space.What if the same applies to humans?
Incompatibilism and the illusion of free will
You probably know very well an argument made by average 20 year old college students (and also Sam Harris) about the non-existence of free will:
- The world is physical;
- Every physical process has a cause;
- Our bodies are physical;
- Therefore our choices are determined and there is no space for anything like free will.
In other words, every decision we make is determined — and the Laplace’s demon would know everything about our future choices — which would mean that we are not in control over what we do — or there is some random element, a chance, that… would mean the same (if the cause is random, we are not in control of the action).
As an addition, there is a famous Benjamin Libet experiment with Bereitschaftspotential which suggests that our body decides what we’re about to do way before we make a conscious decision. Therefore — say incompatibilists — free will is an illusion that is made up by our brains to rationalize something that is already done. So there is no such thing as the ability to “do otherwise”.
How free will is really free
Well, we all have the feeling that we do have freedom of our choices. Deterministic incompatibilism sees it as an illusion that is created by our brains. However, as Christian Kupke points out, if free will was an illusion, there has to be a neural correlate of it in our brain — and we can’t find such a thing. If we stick to deterministic incompatibilism, we are not really able to explain why we need the “illusion” of freedom — just like we can’t explain the phenomenon of meaning. Why would we need it, if we could function without it and survive?
Let’s take a look at how the human decision process works. As we mentioned before, the human mind is able to create abstractions, intentional objects, that play the main role in organizing the world. Consciousness is also based on an abstraction of ourselves — the identity that does not change with the passage of time. We can change our character, but if we remember some choices we made in the past, we are sure that they were made by the same entity, “the self”.
The same way we look in the future — when making a decision we are creating an abstraction of ourselves that does not exist yet. So I can decide if I should have a steak or a salad for lunch based on a biological impulse that my body sends to me, and that is pretty much a question of material determinism. But when I make a choice I don’t only have the material impulse, but a whole concept of what this choice will cause. Can I afford the steak? Is a salad more healthy for me? How would I feel after eating a steak? And this is just about the question of “what should I have for lunch”.
With moral questions there is also another element of this mental journey to the future: the theory of mind. We not only perceive the direct consequences of what we are planning to do, but we are also aware that others will see us and judge us. This can also apply to the choice of what to eat if it becomes a moral question (e.g. whether or not to eat meat on Friday or if it’s okay to kill animals for food).
Intentional objects, such as values, also create biological stimuli (e.g. the vision of us doing something good releases hormones that cause pleasure) so from the deterministic point of view how these intentional objects impact our bodies would also be a part of the causal order. However, based on what we know right now, we can’t simply reduce values and other abstractions just to the activity of certain parts of our brains and bodies. Different people react differently to all kinds of stimuli, including values and ideas. We can at least say that this process is incomputable, so it does have nothing to do with classic Laplace demon determinism.
