Free Will

I’ve got to admit, I don’t understand all of the fuss about this concept. It seems like one of those topics philosophers like to argue about ad infinitum without any progress. Among non-philosophers, it seems there is a tendency to hastily reduce this to a binary distinction, while ignoring the fact that “Freedom” itself is a gradient and the “Will” is simply our capacity to interact with our environment in a non-arbitrary manner; it is a brute fact about ourselves. To me this is not a matter of classification, rather a question of degree under various conditions. If we think about the semantics of the lexemes involved in the phrase, it should be clear that “Freedom” is a type of degree-term or grade-term, with respect to a set of operational criteria. It can also be thought of as a rank-term, where each rank corresponds to an interval with gradeability from lower to higher; when the limit is reached there is a discontinuous jump into the higher rank. My whole point is that we ought to think of this concept in terms of a gradation; once we do this, we can have more productive conversations about the extent to which our “Will” is constrained, given a set of circumstances. This is what I have in mind:

My idea here is that on the lower left, we have a set of states characterizing an agents freedom. Depending on a range of constraints, other agents could be located in the top right region. Obviously the relations and categorizations will be loose and fuzzy. I am also by no means, saying "this is it". I just mean to say a conceptualization resembling this would allow us a wider range of flexibility with our discourse about the topic.

I think a Systems Science perspective is useful when analyzing this concept. The discipline provides concepts that can help us differentiate between systems that have the property of “being free” and systems with varying degrees of teleology. It should be clear that we are not automaton; I merely take this for granted, it does not need to be argued for. Biological organisms are not finite state machines, despite being able to model them as such; one should never confuse the map with the territory. Biological organisms vary with respect to the degree in which their behavior can be conceived of as “deterministic”, but human behavior is inherently stochastic from the observers perspective. This is not to say that the behavior is inherently random; rather from the stand point of an observer, the set of all possible behavior in a given situation cannot be known with any level of reliability for extended periods of time without a sufficient degree of control (and even then, this can lead to spontaneous responses that cannot be known a-priori). Long run behavior is non-ergodic; the “time average” of one system is not equal to the ensemble average across every system. In simple terms, what I am saying is that human behavior looks like Brownian motion at times due to the very fact of our Will being non-deterministic. From a systems modeling perspective, there are no known exogenous factors that cause system dynamics to deviate in a predictable manner over long periods of time. Human behavior does not tend towards an equilibrium; rather there are rapid regime shifts at the macro level that occur from fundamental unstable divergent properties at the micro level. I bring all of this up because “lack of free will” discussions seem to center around certain levels of predictability of human behavior; if we take a closer look we find that human individuals and collectives can be incredibly unpredictable given a set of predictors, in non-arbitrary ways (non-arbitrary and deterministic are not synonyms).

What distinguishes us from automatons? Humans have various properties that can collectively be thought of as “The Will”. We have the property of drivenness; we are able to form and reformulate desires with respect to a set of information and stimuli. We are sentient; in other words, we can experience qualia. We are capable of goal formation and goal setting, given our qualia and desires. We are fundamentally agents who need not respond mechanistically to normative standards of rationality, compounding the predictability problem even more. All of this implies we have consciousness and intentionality; our behavior is directed to ends and we are aware of it. All of this means we have a degree of decision making capacity. This provides us with certain capabilities, such as the ability to arbitrarily interact and fundamentally change the environment we operate in to achieve these objectives. We can engage with complex texts and discover an entirely new interpretation that reorients these goals in light of these understandings. I take all of this to mean what we refer to when we utter “so and so has a Will”. Of course someone will respond to this by saying “But goals can be influenced by external pressure, or our drive can be subject to neurobiological failures”; but this person misses the point and is assuming that if we don’t have absolute control over every possible scenario where something external intervenes, then we somehow don’t have a Will and that its not free. Remember, I am establishing a minimal set of criteria; after that we can refer to all of the situations where this set can be constrained or interfered with. And at a more basic level, something that is causally unaffected by its environment is not "free". In order to act, an entity must be in relation to something else, which opens the possibility of causal influence. These relations are not deterministic, on my view determinism is obviously false or at the very least, has never been substantiated. It must be possible for an agent to be causally influenced by its environment, in order to even speak coherently about Free Will. What distinguishes an agent from an automata are the complex mechanisms inherent within the system that allow the agent to respond non-deterministically to external information. Biological life would simply not make sense on a reductive view of determinism. There would be no reason for entities to evolve. The very nature of life itself implies non-determinism. The fact that we are radically embodied and contextually bound provides the possibility for free agents to interact with such an environment.

Freedom can be contrasted with Controllability. Suppose someone makes the claim that the sub-conscious determines (completely controls) conscious mental states. This would effectively be equivalent to stating that the conscious merely collapses or reduces to the sub-conscious. But as I noted above, I qualify an entity who has a “Will” as one who has the property of consciousness; so this remark is simply inapplicable because it does not refer to entities that can have a Will. By stating that the sub-conscious “determines” conscious mental states is to disagree, ontologically, with what a human is. The correct statement would be “the sub-conscious influences the conscious non-deterministically”; which would simply be a restatement of the fact that we have at least some degree of free will. But at this point, we are quarreling about the degree of freedom, or where we are on the freedom gradient; and whether it constitutes “true” freedom. So obviously, the question of whether we have a Will seems non-contentious; now we must decide whether it is entirely constrained by means of an external controller.

Another common retort might be that since the brain is algorithmic, running on a structure similar to a distributed graph, that the brain must be procedural and subsequently “deterministic”. This again, is simply a confusion; the interlocutor is merely describing how brain states come about, but a description of the inner functioning does not negate the idea that there is a Will, that the system as a whole uses these mechanisms to act freely within bounds on its environment. I think people think that to have a “Will” there must be something immaterial, in addition to, the processes associated with brain activity. This is a mistake; consciousness is an emergent property of this complex system. Saying that there are underlying processes is irrelevant. Postulating an invisible immaterial entity does not solve the problem; in fact I think it complicates the situation and is motivated by metaphysical folk theories that were generated during an era of ignorance. Likewise, someone might be tempted to claim that the existence of any constraint on this Will at all amounts to non-freedom; but again this is ridiculous. Constraints are enabling, and are necessary for an entity to persist. Consider that we necessarily live within a system. This is defined as a 6-tuple S = {C,N,I,B,K,H}, where C = components and their properties, N = network description of the components, I = description of the environment in which the system operates, B = Boundary conditions, K = set of interaction and state transition rules between components, and H = history of the various states (memory). The very fact that we live in a system, and are composed of systems, entails the existence of arbitrary relations that necessitate constraints between components and the system as a whole. Relation implies constraint; but constraint does not imply controllability. Relations also imply the possibility of new potentialities that the system might not have had access to at prior states. This could manifest in a new set of choices for an agent. Relations can also evolve in response to past decisions, but non-deterministically. See No entailing laws, but enablement in the evolution of the biosphere, for a richer description of indeterminism. The fact that we are situated within a set of relations and are composed of a set of process does not undermine the idea of free will unless "free will" is defined incoherently, which I believe it has been historically. All it means, more or less, is that there is no free lunch and that we are situated in a context more or less outside of our control, but with innate capabilities and dispositions to alter our surroundings to our liking.

A bit more on what I've mentioned before, about how the definition of this concept has been radically insufficient and has caused more confusion than anything. At a basic level, many people take "free will" to mean "ability to have acted otherwise in a given situation". This intuitive notion can be undermined by situations such as, falling off a cliff. One might say that we are determined by the physical laws to come smashing to the ground. Therefore, we are determined. This is obviously ridiculous, because presumably we had control prior to that situation, and presumably others can intervene to save us, given some technical constraints that are undetermined. So, there are local regions of determinism, but this does not generalize to global possibility spaces. But obviously, we aren't free in all situations. You can imagine scenarios where you may be fated to make a decision based on a historical chain of decisions, all pointing to a single node in a decision tree. Does this mean we are determined, globally? No, this is also ridiculous. This is why I've stressed the fact that these are gradable concepts and that people have been confused. This is my point, at the conceptual level, philosophers have not established coherency among the terms of discussion. Lets analyze a very stupid conception of free will that has plagued us for millennia. Libertarian free will is a philosophical concept within the broader debate on free will and determinism. It posits that human beings possess genuine free will, meaning they can make choices that are not determined by prior causes or external constraints. This view contrasts with determinism, which suggests that all events, including human decisions, are causally determined by preceding events. Both of these extremes simply miss the point, they are not even false. Libertarianism often emphasizes "agent causation," where the individual (the agent) is the originator of their actions. This means that the agent has the power to initiate actions independently of deterministic forces. But this is ridiculous, if actions are not causally influenced by prior causes, they may be random or arbitrary, which undermines the notion of control necessary for free will. The concept is incoherent. Lets take a look visually why its a contradiction, through the use of a causal graph. It's honestly kind of hard to even draw this since its a contradiction. so bear with me. Consider a simple two agent system. Agent A has Libertarian Freedom, while Agent B does not:

But since we are all free agents, B must also have the ability to act without causal influence from its environment:

But this would mean that causal arrows extend from both source nodes:

If A or B have genuine causal power over their environment, this would imply that it is impossible to draw a graph with nodes extending outward to other agents. So the graph would reduce to something with no edges:

This implies that Agents can have no causal influence on their environment. But this contradicts the definition of "Agent Causation". The system reduces to a collection unconnected nodes randomly moving around, since it is impossible to draw a causal graph of a multi-agent system that does not contain nodes with incoming causal arrows. There is one exception: the case where the root node does not respond to exogenous influences, but can effect every successor node. This can be visualized here:

So in order for "Agent causation" to be coherent, there cannot be more than one node in a graph that has causal arrows pointed outward. This implies that it is possible for ONLY ONE agent to have libertarian free will. Therefore, humans as a whole, do not have libertarian free will. It is a contradiction to say that we do. In summary, it is impossible to draw a causal graph of a multi agent system with agents thought to have "Agent Causation". The only way this is possible, is with what I drew above, or a situation where there are no causal arrows emanating from nodes; which would mean that agents cannot causally act on one another, which contradicts the idea of agent causation. So there are two options, described below in these graphs:

You can have a conception of free will that allows for the possibility of agents being causally influenced by their environment. Or you can bite the bullet, which would imply the causal graph looks like this:

In this second scenario, we have a system of agents that look more like random particles in a vacuum. Undetermined, yet completely random. This is absurd to me. Therefore, I think we really ought to reconsider what we mean by the term "Free Will".

Anyways, back to the qualifier “Free” in “Free Will”. The existence of constraints qualifies the degree, rank, and grade of “Freedom” that our will has. But more importantly, the existence of a top-down controller (akin to a control system) sure seems to be equally important. One could say that if our choices were entirely constrained by a controlling mechanism, our “Will” would have the least amount of “Freedom” but would still have the capability of “Freedom” with the non-existence of such restrictions. So “Free Will” is a capability fundamental to conscious agents, allowed by the functioning of biological processes. Is our “Freedom” path-dependent? Yes. Do we make decisions like weak-learners who use simple reward functions? Presumably no. Economists have had quite the trouble reducing choice to utility maximization given a set of preferences and budget constraints. Do we respond to incentives? Yes, but incentives tend to be very unstable and unpredictable. Is the system Controllable? No. All of these considerations add and subtract to where we fall on the “Free” scale. I am not going to argue where we fall on the scale, I just wanted to show that I think its indisputable we “have” a “Will” (I quote the word “have” because even phrasing it as a Noun can lead to weird misunderstandings about our fundamental nature. It’s not something that we “acquire” or are “given”, it is simply a brute fact about ourselves) and that there is no know top-down controller necessitating decisions. Constraints will qualify the “Freedom” our “Will” can express, but this is a matter of degree.

It should be clear by now that I view Free Will, or some approximation of the concept, as fundamental to humanity. However, I think the concept as understood historically and philosophically obscures what it means for us to have agency. Libertarian notions of Free Will are inherently connected with Theism. As I noted above, this concept is a contradiction in terms. But lets look at the one scenario where Agent Causation is possible: the situation where a maximum of one agent can have this sort of freedom. I also want to connect this to the notion of a Controller.

Classical Theism assumes a Tri-Omni god: Omnipotent, Omniscient, and Omnibenevolent. The only agent who could possibly have libertarian freedom would be a god, since this entity can exert causal influence on its creation but the reverse is not possible. If this entity plans to act, it does so without error. Its plans are executed as intended. Therefore, its causal interventions are deterministic. Furthermore, an entity with unconstrained ability to act on the environment it has created, combined with omniscient knowledge, cannot be surprised by anything within its creation, since it would have anticipated it in advance. In other words, nothing can deviate from this plan, unless it were not omniscient or lacked the ability to execute its intentions. Given this characterization, its hard to see how this agent isn't a deterministic controller. This has always been intuitive to me, even prior to familiarizing myself with philosophy of religion. It turns out, there is a formalized argument noticing this tension (in my view, contradiction) between a libertarian notion of free will, and the existence of a deity described by classical theism. Before I present the argument, consider this parallel.

Many theists assert that Atheism entails determinism, usually by conflating causal determinism with Atheism based on a misunderstanding of physics and non-belief (the fact that Atheism is not a world-view). According to Laplace's demon, the main view of classical determinism, if someone (the demon) knows the precise location and momentum of every atom in the universe, their past and future values for any given time are entailed and can be calculated from physical laws governing the universe. Make note of the condition "to know"; Laplace is asserting that the only thing limiting our ability to see this determinism, is the fact that we are not omniscient. In other words, if we knew the relevant information, we could retrodict and predict the entire state space of the physical universe. This form of determinism assumes that:

The Universe is Governed by Physical Laws: Nature operates through unchanging physical laws that determine the behavior of every system.
Complete Knowledge Implies Predictability: If all information about a system is known, including initial conditions and governing laws, future and past states can theoretically be computed.
Causal Closure: Every physical event has a physical cause.
Absence of Randomness: At least in classical mechanics, no randomness or indeterminacy exists; everything unfolds in a predictable chain of cause and effect.
Sufficient Information and Computational Power: The hypothetical intellect would require infinite computational power and perfect knowledge of initial conditions, which is practically impossible (with limited human cognitive resources).

Notice how assumptions 2-5 seem to perfectly coincide with an entity characterized by classical theism. Nothing happens outside of an omnipotent entities control and everything is known by an omniscient entity, therefore Laplace's demon is the entity described by classical theism, which means classical determinism is entailed by classical theism. Libertarian free will is not only a contradiction in most circumstances, but the one situation where it is applicable, entails determinism for every other agent. The assumption that Libertarian Free Will is incompatible with Atheism, simply collapses on itself because classical theism is also incompatible with this notion of free will. As I've argued above, I think this is due to problematic conceptions of the terms "free" and "will". Once we realize this problem, it becomes clear (to me) that free will can only be compatible with the non-existence of a deity. Even if the classical theist rejects Libertarian Free Will, the characterization of a classical entity implies Laplace's demon. The only possible solution is to consider alternative notions of theism; such as Open Theism. This critique applies to all forms of traditional theism, not just Calvinism.

This brings me to the argument that I wanted to present earlier. This falls under the class of philosophical problems known as Fatalism, the idea that situations or events were somehow inevitable. I touched on this concept indirectly earlier with the idea of local determinism. Obviously, I reject global fatalism. In the context of classical theism, many people have noted the problems of divine Foreknowledge and Free Will, which I see to be crucially problematic for many conceptions of Theism. The argument for theological fatalism is presented below informally, then formally:

For any future act you will perform, if some being infallibly believed in the past that the act would occur, there is nothing you can do now about the fact that he believed what he believed since nobody has any control over past events; nor can you make him mistaken in his belief, given that he is infallible. Therefore, there is nothing you can do now about the fact that he believed in a way that cannot be mistaken that you would do what you will do. But if so, you cannot do otherwise than what he believed you would do. And if you cannot do otherwise, you will not perform the act freely.

Yesterday God infallibly believed T. [Supposition of infallible foreknowledge]
If E occurred in the past, it is now-necessary that E occurred then. [Principle of the Necessity of the Past]
It is now-necessary that yesterday God believed T. [1, 2]
Necessarily, if yesterday God believed T, then T. [Definition of “infallibility”]
If p is now-necessary, and necessarily (p → q), then q is now-necessary. [Transfer of Necessity Principle]
So it is now-necessary that T. [3,4,5]
If it is now-necessary that T, then you cannot do otherwise than answer the telephone tomorrow at 9 am. [Definition of “necessary”]
Therefore, you cannot do otherwise than answer the telephone tomorrow at 9 am. [6, 7]
If you cannot do otherwise when you do an act, you do not act freely. [Principle of Alternate Possibilities]
Therefore, when you answer the telephone tomorrow at 9 am, you will not do it freely. [8, 9]

The main idea is that foreknowledge and omniscience are incompatible with free will. As I've alluded to above, I think omnipotence further complicates this problem, although its not formally acknowledged in this particular argument.

I'll note that there have been responses to this argument (see the SEP article linked above). Many of them fall under the category of "compatibilism", which is the idea that determinism and freedom are somehow compatible. Many atheists are compatibilists as well. They think that the universe is "determined" but free agents can still act freely. It should be obvious that I hold to incompatibilism. I think that, if determinism were true, then it would be hard to see how we are free (although maybe there would be a convincing argument out there not yet discovered, who knows). But I think determinism is obviously false and we need to really rethink our notions of causality and agency. I also think that every compatibilist argument presented in the context of theological fatalism are deeply unconvincing and that if theism were true, determinism is true. My approach is somewhat of a Moorean Shift; it is obvious determinism is false, therefore theism (as traditional conceived) must be false. This is a "Free will argument for the nonexistence of God", which I think is quite plausible but needs more elaboration. I'll give one example of a response to theological fatalism that is ridiculously uncompelling. Many theists will say the following:

God knows what we will choose, but so long as his knowledge is not a cause of our choice our will is free within its constrained perspective.

This is one of those responses where we must ask ourselves "did this person even read the argument?". Yes, knowledge does not cause the choice, but god knowing what we will choose, still implies we could not choose to do otherwise. This simply misses the point of the entire argument. Also, earlier I mentioned that omnipotence complicates the issue. I think its apparent here because god engineered the possibility space in this given scenario, but since god has a will of his own that deterministically manifests, the set of options must have been illusory.

The topic of free will is interesting because I think whatever conclusion you draw depends on how you conceptualize modality, causality, and time, which are interesting topics by their very nature. I think that this contentious topic is a function of our misunderstanding of these more primitive concepts. Part of my motivation for writing this was due to peoples misunderstanding of "scientific laws" being somehow binding across all levels, upwards through the system of systems characterizing reality. One thing I didn't mention was how free will overlaps with adjacent philosophical topics such as moral responsibility. I think many theists very unfairly mischaracterize non-belief by attributing determinism, and subsequently a lack of moral accountability, to be some inherent quality of being an atheist. I mentioned in previous posts this fear mongering that serves to retain believers and leads to demonization of non-believers. So despite the Free Will debate being somewhat of a philosophical language game, it has nevertheless initiated interesting lines of inquiry concerning some our most fundamental presumptions. I personally like the topic because I am interested in Decision Theory, and more broadly, how people make choices and rationality.

Below are some links to resources I've engaged with:

General Resources:

Debates & Discussions

Search This Blog

CriticalThinkingAcademy

Free Will

Comments

Post a Comment

Popular posts from this blog

The Nature of Agnosticism Part 5

Core Concepts in Economics: Fundamentals

The Nature of Agnosticism: Part 4.4