This Isn’t the Free Will You’re Looking For

by Steve Neumann

In Star Wars: A New Hope, Luke Skywalker and Obi-Wan Kenobi are traveling into the most “wretched hive of scum and villainy” that is the city of Mos Eisley in order to find a smuggler who can hide them, R2-D2, and C-3PO from the Empire. At one point, they are stopped at a checkpoint by Imperial Stormtroopers who are looking to recover the stolen technical plans of the Death Star that R2-D2 is carrying. Obi-Wan uses a Jedi mind trick to divert their attention, telling them “these aren’t the droids you’re looking for.” 

In a paper entitled “Freedom and Control,” philosophers Stephen Mumford and Rani Lill Anjum [1] have attempted to capture free will for those who are libertarian about it — those who hold that, since humans do have free will, determinism must be false. But what they end up doing, in my view, is arguing for a new species of compatibilism; except in this case, free will is compatible with causation (properly understood) and not determinism. And to paraphrase old Ben Kenobi, I don’t think this is the free will libertarians are looking for.

Mumford and Anjum argue that the libertarian can have free will “without requiring that agents step outside of the causal laws.” If libertarians could achieve this, it’d be the Holy Grail of metaphysics. But I think the authors implicitly recognize that such a goal might not be possible when they write that they can “supply a variety of libertarianism worth having,” echoing Daniel Dennett. [Italics mine.]

Mumford and Anjum claim to have secured both freedom and control — or what are called Alternate Possibilities (the ability to have done otherwise) and Ultimate Authorship (one is the ultimate author of one’s own decisions that lead to actions, or “causal responsibility”), respectively — by arguing for a view of causation that is somewhere between “necessity and pure possibility,” where possibility can be understood as akin to randomness [2]. They maintain that a strict dichotomy between necessity and randomness is a false one, and they note that there is a third possibility that goes by the  name of “causal dispositionalism.” In other words, they deny both causal determinism and causal necessitation. [3]

Here’s a very simplistic diagram of the basic background to the problem:

Fixed Past => Present Possibilities/Choices => Future

Now, surely most of us can agree that the past is fixed — we can’t change it; what happened, happened. And, all of us experience the fact that, at any given moment, we seem to have a near limitless number of possibilities available to us when making a decision. The problem, however, comes in when we try and figure out what to make of the future — is it fixed, or is it open? Do past events completely cause or determine our choices in the present, even though we may feel like we have many possibilities available to us?

Mumford and Anjum argue, echoing Aristotle, that “a cause can be thought of as something that tends towards an effect of a certain kind,” that it is disposed toward a certain effect — hence the name. With this definition, they further claim that we can separate the ideas of causal production and causal necessitation by “drawing attention to the possibility of a particular variety of interference that applies to all natural causal processes.” In other words, causes don’t necessarily produce their signature effects. This idea of “interference” is where they hope to help the libertarian locate free will by showing that alternate possibilities really exist. 

For purposes of my critique, I’ll use one of the examples Mumford and Anjum cite: philosopher J.L. Austin’s famous thought experiment of a golfer sinking a putt. They note that “the ball would not have sunk had a gust of wind come along just as it neared the hole, or a squirrel might have jumped on the ball, or a twig might have deflected it. These alternate possibilities are all real.” The wind, the squirrel, and the twig are all examples of the interference they refer to in the context of causal dispositionalism. 

In the case of Austin’s golfer, it’s easy to see how alternate possibilities could be secured: it’s true that before he attempts the putt it seems as though there are many possible outcomes: a sudden gust of wind could come along just as the ball nears the hole; a squirrel could run across the green, knocking the ball off its trajectory; or the golfer himself could have a heart attack just as he swings his club. These are all very real possibilities — interferers, in Mumford and Anjum’s lexicon. Additionally, the golfer is employing his best causal powers in his attempt to sink the putt, and these powers are disposed toward the outcome of the golf ball falling into the hole; so they also conclude that the golfer is the ultimate author of his actions. They summarize their position by saying that “in exercising our causal powers, we exercise our free agency.”

But this doesn’t seem quite right. I’m ready to agree that we exercise our agency by exercising our causal powers — where agency means the ability to voluntarily act upon our own reasons and motives — but to me that still doesn’t mean we have free agency, at least not in the libertarian sense.

At the moment of an action, there may indeed be several possibilities available to an agent; however, these possibilities aren’t freely chosen in the sense the libertarian wants: for instance, each of these possibilities is itself subject to its own causal chain which no agent can step outside of — causal powers are still comprised of causes, deterministic or not. And if the interferers that contribute to these alternate possibilities originate outside the agent (a malicious squirrel), thwarting the agent’s expressed desire (of sinking a putt), then the agent has no ultimate control. And even if the interferers arise within the agent’s own mind, either as a counter-desire to his initial one, or as an external influence that works in his mind to alter the initial desire, then where is the freedom in that? The case of the external interferers doesn’t seem to pose much of a problem — I think most people can accept that a sudden gust of wind altering the trajectory of the golfer’s putt is something that is completely out of his control (and this is one of Mumford and Anjum’s arguments against causal necessitarianism). So let’s focus on the internal interferers. 

Let’s say Austin’s golfer decides to position his thumb a certain way on his club just as he’s about to make his putt. However, right before he swings the club, he has the thought to move his thumb into a different position; but what is the cause of that thought? The thought to change his thumb position simply occurred to him. Where is the freedom in that? The libertarian assertion that, even though our golfer can’t control the thoughts that come into his mind, he still has control of his actions once the thoughts occur, doesn’t seem to cut it — the golfer’s decision to change his thumb position still wasn’t freely chosen in the sense libertarians want. He may have had a subsequent thought that approved of the thought to change his thumb position, but then one can ask where that thought came from, and so on ad infinitum. 

I think the line of reasoning that Mumford and Anjum pursue ultimately fails to secure the kind of freedom libertarians believe is possible — and which they really want. I think that to say there is “ample space for causation to occupy between necessity and pure contingency” makes sense, and it shows that an agent has alternative possibilities in the face of a decision. And the argument that one is the author of one’s own actions because one is the causal producer of the effects one desires is an accurate and workable conception of agency. But one cannot be the ultimate author of one’s actions, even if causal necessity is false, because the alternatives — dispositionalism and pure contingency — permit one to be only the proximate author of one’s actions. Compatibilists about free will are okay with this, but not the libertarians.

Pure contingency (randomness), I think everyone agrees, doesn’t allow one to be the author of one’s actions, so let’s return to the issue of causal dispositionalism. Here, even if one is acting according to one’s own singular character, to one’s own developed reasons and motives — those things that dispose the agent toward certain effects — that character, and those reasons and motives, still have their own causal history outside of which no agent can step. The reasons and motives that make up the agent’s character didn’t pull themselves up into existence by the hair, out of the swamps of nothingness, to paraphrase Nietzsche’s take on an agent being causa sui. The agent hasn’t freely chosen those reasons and motives, even if she has come to approve of them and take “ownership” of them.

For a libertarian to feel that she has free will, she needs to feel that she has real alternate possibilities available to her when making decisions, and that she is the author or originator of her actions in a meaningful way. But the biggest obstacle to this has always been the perceived failure of libertarian models of agency to explain how an agent could have done otherwise in exactly the same conditions. Some have argued that the original set of circumstances will always obtain, every time you “roll back the film” of the action in question. Others have argued that, due to the inherent unpredictability of the universe, the original set of circumstances may not always obtain for the action in question. But even in the latter case, even if quantum indeterminacy obtains every time we roll back the film, ensuring a slightly different result, the agent can’t claim to have “freedom” then either: an agent can’t be the author or originator of a completely random event, at least not in the sense of having willed it.

In the concluding section of their paper, Mumford and Anjum write that the principle of Alternate Possibilities was “threatened by a world of necessity,” and that the principle of Ultimate Authorship was “threatened by a world of pure contingency,” and that the causal dispositionalist framework they use in their argument “provides a metaphysical basis highly conducive to the libertarian’s needs: one in which the agent has both freedom and control.” But unfortunately I have to resort to a Clintonesque parsing of the word “freedom” and say that what they have actually secured is a compatibilist’s freedom, not a libertarian’s. They correctly note that the “more empowered an agent is, the more freedom they gain,” but this is the kind of freedom that is conducive to the personal project of identifying, developing, and expanding one’s causal powers so that one can attain greater degrees of freedom in the sense of achieving real possibilities of life. It’s a valuable freedom, no doubt, and in reality the only one we have — and, in light of that, the only one worth having. 

________

[1] I’ve invited Mumford and Anjum to comment on this post, if they have the time. Hopefully they can set me straight if I’ve misread their paper!

[2] The actual formula used by the authors for the principle of Alternate Possibilities is stated as “for any free agent x, and action A performed by x in circumstances C at time t, then there was another action A’, where A does not equal A’, such that x could have performed A’ at t and not A.” Though they leave out “circumstances C” in the second clause, I’m assuming it is included for purposes of my analysis.

[3] I haven’t read their book on causal dispositionalism, so I’m basing my critique on their summarized exposition of it in this current paper.