Are hierarchical theories of freedom and responsibility plausible?

In order to be responsible for your actions, you must be free. Or so it is commonly believed. But what exactly does it mean to be free? One popular view holds that freedom consists in the ability to do otherwise. That is to say: the ability to choose among alternative possible futures. This popular view runs into a host of problems. The obvious one being that it is inconsistent with causal determinism.

This has led several authors to propose alternative hierarchical theories of freedom. According to these theories, an action is free when it is consistent with an agent’s higher-order, reflective desires. The idea is that sometimes we have impulsive, non-reflective desires that are not consistent with the kinds of people we really want (or believe) ourselves to be. I, for example, currently desire a piece of cake. But I have also, in my more reflective moments, committed to losing weight because I want to be a skinny person (note: this is just a hypothetical). Consequently, acting on my impulsive desire for cake would be inconsistent with my higher-order preference for being skinny. That would be the essence of unfreedom.

Hierarchical theories of freedom have many attractive features. They are consistent with determinism, and speak to the core belief that in order for an action to free it must belong to us in some respect. But do they provide a compelling account of responsibility? In his book, Against Moral Responsibility, Bruce Waller argues that they don’t. Indeed, he argues that the overwhelming belief that freedom and moral responsibility are connected has led people to propose deeply flawed theories of freedom and responsibility. The hierarchical theory is just one particularly good example of this.

In this post, I want to review Waller’s main arguments. I do so largely as an attempt to better understand his critique. I have, in the past, endorsed hierarchical theories, but have recently become more sceptical. Waller’s critique proceeds in three stages, each one looking at a variant of the hierarchical theory from a different theorist — Harry Frankfurt, Gerald Dworkin and Susan Wolf, respectively. I’ll sketch each of these three stages in what follows.


1. Frankfurt’s Theory and the Implausibility of the Hierarchical Approach
Harry Frankfurt’s 1971 article, “Freedom of the will and the concept of a person”, is perhaps the classic work on hierarchical theories of freedom. It proposes the simplest, and arguably most compelling, of the hierarchical theories. This is the one I laid out in the introduction. It claims that freedom consists simply in doing whatever is consistent with one’s second-order preferences. And what exactly is the difference between a first and second (or higher)-order preference? The answer is roughly as follows:

First-Order Desire: Is expressible in the form “A wants to X”, where “X” is some particular action (like eating cake)

Second-Order Desire: Is expressible in the form “A wants to X”, where “X” is some particular first-order desire (like wanting to want to eat cake, or, in my case, wanting not to want to eat cake).

So any particular action is free when the desire motivating the action is endorsed by a higher order preference to want to have that desire.

The problem with this simple theory is that it appears to have troubling implications. It implies that certain people who we would not ordinarily classify as being free are in fact free. In particular, it implies that people who are gripped by compulsive desires are, sometimes, free. Frankfurt embraces this implication when he distinguishes between two kinds of drug addict. Ordinarily, we would be inclined to say that the drug addict is not free: she is controlled by her first-order desires to take a drug. But Frankfurt says this is not always true. There are wanton addicts and willing addicts. Wanton addicts follow their first-order desires, even when they are out of line with what they really want to want (a good job, a stable family life etc.). Willing addicts fully endorse their first-order desire for drugs: they want to want them. They are truly free (and, by implication, responsible for what they do).

Waller argues that this is absurd, particularly when we bear in mind the typical history of the willing addict. Consider three counterexample:

Willing Addict: Peter starts taking drugs in college. He initially believes himself to be in control of his desire, saying “I can quit anytime”. Later, he finds himself trapped in a drug addiction he despises. He tries to get out of it but instead he slides deeper and deeper into difficulties. He loses his family and friends, destroys his career, and suffers from numerous psychological and physical problems. In the end, nothing of his old life is left. At this stage, he has an epiphany: since nothing of that old life is left, he has no reason to despise what he has become. He then embraces his addiction. He wants to want the drugs. He becomes a willing addict.

Willing Slave: Jamal is a fierce, independent warrior. He his captured by slavers and transported to a plantation in the Caribbean. While there, he his “whipped, branded, and abused”. He is forced to work against his will. In the beginning, he maintains his commitment to freedom, striking back at his slave masters whenever he gets the chance. But, after many years, he gives up. His spirit is broken. He embraces his conditions. He becomes a willing and happy slave.

Willing Convert: Eve is a strong, independent young woman. She longs for an education and career of her own. Unfortunately, she has been born into a strict, religious community. In that community, women are expected to be meek and compliant, to accept male authority, to remain uneducated, and maintain a subservient societal role. Eve rejects those values and “insists that she be respected as fully equal to anyone else”. But after years of “failure, condemnation, and psychological and physical abuse”, she breaks down. She starts to accept the subservient role. She becomes a willing convert.

In each of these cases, the individuals in question meet the conditions set down by Frankfurt. In the end, each of them reflectively approves of their first-order desires. But surely we would not say that any of them are free? Indeed, they arguably epitomise unfreedom. This suggests that Frankfurt’s simplistic version of hierarchical freedom is deeply flawed. The question is whether the hierarchical approach can then be salvaged.


2. Dworkin and the Right Causal Pathway Account
If you look at the three counterexamples just given, you’ll notice a common theme. They each involve people who have to embrace their position in life via a certain kind of causal pathway. Namely, a causal pathway involving their deprivation, abuse or coercion. They want what they now have, but only because circumstances left them with no other viable options. This cannot be freedom.

But this directs our attention to a possible escape route for proponents of the hierarchical approach to freedom. Instead of arguing that freedom is simply about wanting what you want, couldn’t they argue that it is about that and coming to that realisation through the right causal pathway?

This is exactly what Gerald Dworkin claims in his 1988 book The Theory and Practice of Autonomy. He argues that one’s higher order evaluations need to meet the condition of procedural independence. Very roughly, this means that one’s higher-order evaluations are free from manipulation and coercion. That they are arrived at through appropriate education and access to the right kinds of information. They do arise simply because one is beaten, abused and cajoled into accepting one’s lot in life.

This is an intuitively attractive idea. We all have the sense that certain desires are arrived at via improper causal pathways, and certain others are not. If Eve decided she wanted to live the life of subservience after having received an education and being exposed to the mainstream, secular way of life, then we might view her differently. The fact that she didn’t and was never even given that opportunity is crucial to her lack of freedom.

But Dworkin’s solution raises new problems. Waller highlights two of them. The first is that it effectively does away with the hierarchical component of the theory. If what matters is that you arrive at your desires through the right kind of deliberation, then that’s all that matters: consistency with higher-order desires doesn’t seem like an necessary addition. The second is that it is difficult to know what counts as the right causal pathway. Manipulation and coercion by others is one thing, but what about more subtle forms of manipulation? We are all “manipulated” by our genes, culture, education and social setting. Countless studies confirm this fact. Do these count, and if not why not? Waller gives the example of a willing gambler, who came by his addiction due to a fortunate run of luck the first time he visited a casino. Did he arrive at his compulsion through the right causal pathway? If not, then any number of “fortuitous contingencies” would seem to undermine freedom. The number of truly free actions would be vastly diminished. Maybe that’s something we are willing to accept, but we should acknowledge it as a potential consequence of Dworkin’s theory nonetheless.

Dworkin proposes a test of his own. He says that one way we can know if a desire is arrived at in the right way is if the individual in question would reflectively approve of the process whereby they arrived at that desire. Thus, I can say that I arrived at my desire to be skinny through careful deliberation about the person I would like to be; and I approve of that process of arriving at that desire. I’ve come to this position in the right way. The danger with this test is that many people who have been manipulated or coerced into a state of acceptance are likely to reflectively endorse the process whereby they arrived at that state. So, for example, Eve may well approve of her frustrations and denials once she has “come to see the light”. She may thank her community for helping her to see the errors of her ways. That doesn’t make her free.


3. Wolf’s Perfect Rationality Account of Freedom
There may, however, be one causal pathway that leads to freedom. This is the one advocated by Susan Wolf. According to Wolf, the only way to be truly free — i.e. free from the kinds of manipulations and coercions we worried about above — is to track the True and the Good. In other words, to desire what is right for the right reasons.

This view has its origins in religious, predominantly Christian, philosophy (though the Christians adopted it from the Greeks). The idea, as Waller describes it, is that:

True freedom is living in accordance with one’s true nature (as a rational being); genuine freedom can be realized only through accurate pursuit of the True; real freedom means living in accordance with the way God designed you; true freedom is found in perfect obedience to God. 

(Waller, 2011, pp 66-67)

Wolf adapts this classic ideal by replacing obedience to God with obedience to reason. One behaves freely when one believes what is true, desires what is right, and does so because one has access to the right information and can process it appropriately. In that case, you are not being surreptitiously manipulated into your desires, and there are no subtle, undetectable, genetic or environmental quirks influencing what you do. You are simply being guided by the light of reason.

There is some irony to all this. If Wolf is right, then freedom does not consist in the ability to jump tracks and to do otherwise; instead, it consists in the ability to follow the right track (note: if there are many things that are “right”, there may still be several tracks that one can follow; nevertheless, there is a much narrower set of right tracks than is typically thought).

I have to say, I find Wolf’s account somewhat attractive. I don’t know if I would call it a theory of responsibility or freedom, per se, but I do think it addresses the worries we have about manipulations and other causal influences. The obvious problem with Wolf’s account is that humans routinely and systematically fall short of such perfect rationality. In addition, it may challenge the traditional conception of responsibility for one’s actions. This doesn’t mean she’s wrong, of course, it just means this sort of freedom is alien to human beings. (Waller, I should note, also thinks that the ability to jump track is valuable and neglected in Wolf’s account).

So what should we conclude? I’m not sure. There is much more to the literature on freedom and responsibility. There are several other variants on the hierarchical/right causal pathway theme, and there is a veritable cottage-industry of academic work on manipulations and how they may, or may not, undermine freedom. Nevertheless, I like Waller’s simple criticisms. I think they embody a robust commonsense. I think he is right to say that Frankfurt’s approach is flawed, and that identifying the right causal pathway is extremely difficult (unless we resort to Wolf’s extreme). Philosophical sophistication is all well and good, but sometimes you need that kind of commonsense critique.