Farahany’s Perspective on Free Will, Freedom of Choice, and Freedom of Action

Farahany builds her definitions of freedom and free will by following the work of the philosopher Harry Frankfurt (1969, 1971, 1993). She explains that free will encompasses freedom of choice and freedom of action. Farahany (2016, 61) defines freedom of choice as the “freedom over one’s preferences, desires, and/or dispositions.” She then explains that freedom of action represents “the freedom of intending an action, being able to bring it about without obstacles or impediments, and identifying with the action which results” (Farahany 2016, 61).

Farahany (2016, 61) exemplifies the notions of freedom of choice and freedom of action by describing the behavior of someone who is thinking of eating a cake. Freedom of choice relates to one’s craving, thus freedom over one’s preferences and disposition (for example, one might dislike carrot cake and prefer instead chocolate cake, one might be hungry or not, etc.) Freedom of action, however, differs from freedom of choice in that an individual deliberates through the freedom of action in their decision to actualize their desire. For example, in this case, the person needs to purchase the desired cake, take a fork, cut the cake, and then eat it.

For Farahany, freedom of action is the essential feature that needs to be present in order to establish if an individual is morally responsible for their actions. Farahany reflects on the fact that there are patients unable to perform specific actions through the movement of their body parts (i.e., paralyzed patients who cannot move their arms). However, through the development of BMIs, which transmit information that is processed through specific algorithms connected to artificial arms, these patients are now able to perform the actions they think and desire to do. Their thoughts can be decoded and transmitted to robotic arms so that these patients, despite the paralysis of their limbs, can fulfill their conscious and deliberate intentions to perform their desired actions with their artificial arms. The use of BMIs “demonstrate[s] that action choices are distinct choices with neural representations that can be detected and isolated” (Farahany 2016, 62).

Farahany observes that once brain implants have been fitted, patients will need extensive practice to successfully perform the desired movement of their robotic arms. This is explained by Farahany in the following passage (which is worth reporting in its entirety) describing what has been observed in Tim Hemmes (a paraplegic patient following a car accident) after a BMI chip was implanted into his brain:

After the new chip was implanted in Hemmes’s brain, moving the robotic arm was not as simple as Hemmes thinking, “I want to move my arm.” Instead, he had to discover how to form the specific intention to move the robotic arm and train for weeks to learn how to do so. He began by training on moving a cursor around a screen, quickly discovering that thinking simple thoughts like “move up” or “move down” did not suffice to achieve the actions he desired. Instead, he had to learn a new language, a new way of translating his intention to act into three-dimensional actions. This technological feat makes plain the difference between dispositions (such as inability to move one’s arms), intentions to act (deliberate thought processes), and performance of actions (achieving action desires in two- or three-dimensional space) . . . Hemmes had to learn to create effective brain states to move the robotic arm, making plain that conscious willing of an action [thus the deliberation through the freedom of action] is a necessary cause of an intentional action. His training goes to the core of whether the brain alone controls actions, or whether some conscious “self” exercises control, choice, and movement. Hemmes tried to simply let his “brain” figure out how to move his robotic arm. That approach failed, while Hemmes’s experience of consciously and deliberately training eventually resulted in effective brain states to signal the robotic arm causing the arm to move in accord with Hemmes’s intention. Hemmes identified the resulting movement of the robotic arm as his own action. (Farahany 2016, 62–63)

This patient had to spend time training himself on how to perform a specific action with his artificial arm in order to learn to create specific brain states that were then able to provide the information needed to move the robotic arm in the desired way. This shows that the resulting action (i.e., a specific movement of the artificial arm) was not simply the output of brain mechanisms understood deterministically (e.g., explainable solely in terms of electric impulses and/or release of specific neurotransmitters). Instead, the patient’s actions (performed with his artificial arm) were the result of his free choice; so, it is reasonable to affirm that the actions performed by this paraplegic patient through BMIs were voluntary actions. As we shall see, the experimental studies of BMIs have arguably helped prove that free will (in relation to the performance of specific desired actions) exists. The deliberation that leads one to perform a specific action consists of two steps: (i) first, a decision is made with the freedom of choice (i.e., in my mind, I can choose that I want to eat chocolate cake rather than carrot cake); (ii) second, a decision is made with the freedom of action (i.e., I will move my arm and hand and open my mouth in order to eat the chocolate cake).

According to Farahany, a person should be judged as culpable for their actions not simply if they think about carrying out an evil action (through the deliberation of the freedom of choice that relates to the generation and imagination of all sorts of ideas in our mind) but only if they actually execute the action (through the deliberation of freedom of action, which enables the actual implementation of the desired action). This approach uses a pragmatic understanding of moral responsibility that has been widely accepted in our society, given that a person is culpable not just for thinking of doing something bad but for actually carrying out the premeditated evil action. It is in fact difficult to argue that a person who commits a premeditated crime is not morally responsible for their evil action; however, it is incorrect to judge a person as guilty for simply thinking of carrying out (without ultimately executing) an evil action. This idea is consonant with the teachings of the Catechism of the Catholic Church, which, following the instructions of the Council of Trent, provides some guidelines to define mortal sin. Three different conditions need to be met in order to affirm a person has committed a mortal sin, namely: (i) the object of the sin needs to be grave matter, and there needs to be (ii) full knowledge and (iii) deliberate consent to prove there is “sufficient freedom” in relation to the execution of the sinful action (Catechism of the Catholic Church 1857, 1859; John Paul II 1984, 17 § 9–12).⁴ Thus, the fact that deliberate consent with “sufficient freedom” is an essential requirement for mortal/grave sin agrees with Farahany’s argument that a person can be judged as culpable only if the premeditated evil action is intentionally executed. Farahany in fact argues that only freedom of action (and not freedom of choice) accounts for the actualization of premeditated actions (Farahany 2016, 62–68).

The existence of brain implants that enable human beings to move robotic arms (in order to intentionally perform a specific task) shows that premeditated movements are the result of voluntary decisions that have been deliberated through the freedom of action (and not only through the freedom of choice); a human being can in fact think or dream of performing a specific action, but thinking of doing something does not necessarily mean the person will actually implement their ideas. An extra step is therefore needed to transfer the decision deliberated through the freedom of choice to the freedom of action, and this is demonstrated by the fact that the paraplegic patient needed time and practice to learn how to create specific brain states that were able to move the artificial arm in particular ways.

Let us make one further example to elucidate these issues. Imagine a speaker is delivering a seminar at a conference and an individual in the audience thinks (with their freedom of choice): “What if I suddenly get up, walk to the front of this room, and throw a punch at the speaker to stop him talking.” Before actually doing this, the person needs to think: “I am going to stand up, walk over, and throw a punch.” The final decision to implement these actions is the result of the deliberation of their freedom of action; most likely, this decision will not be implemented, because the person will soon realize that this is actually a bad idea with negative consequences. If we were able to detect and decode their brain activities to see what this person was thinking (through the use of their freedom of choice), we would be unable to accuse the person of having done something wrong for simply thinking of punching the speaker. In other words, we cannot accuse this person of being abusive simply because their freedom of choice enabled them to have this thought. The person can be accused of assault only if the thought of punching the speaker passes their second filter mediated by their freedom of action. Thus, it is the freedom of action, not the freedom of choice, that accounts for moral and criminal responsibility, since it is through the deliberation of the freedom of action that the green light is provided for the individual to actually decide to get up, walk over, and punch the speaker. It is important to reflect on these issues because nowadays, BMIs enable us to “read” and “decode” brain activity through pattern-recognition algorithms (Farahany 2016, 51; Rainey et al. 2020). It would be highly inappropriate to judge culpability following the detection of some brain activity in relation to simply “evil thoughts.”

For these reasons, Farahany is correct to conclude that an individual has acted immorally if their actions were performed in response to deliberation resulting from freedom of action (and not just for possible scenarios envisaged in the human mind through the use of freedom of choice, since these will not necessarily be actualized). Farahany has strong reasons to make this claim, especially after having explained that the use of BMIs on paraplegic patients has enabled us to distinguish between the two types of freedom. Now, the findings reported by researchers when BMIs are used in nonhuman creatures will be detailed, showing the strong similarity to what has been described so far in human patients.

Experiments in Which BMIs Are Used on Monkeys

It has been demonstrated that BMIs can also enable monkeys to move artificial arms (Peterson 2005, 337–39; Farahany 2016, 52; Greenemeier 2008; Ifft et al. 2013). I argue that it is important to reflect on this issue, since this has some important philosophical-theological implications for the understanding of moral theories. What I propose in this section differs from Farahany’s perspective. Farahany explains that human beings have, in contrast to other animals, the ability to form “second-order desires.” This ability, in agreement with Harry Frankfurt’s perspective, imparts the notion of personhood, which represents the distinctive feature of free agents who possess moral responsibility. In particular, Farahany writes that

[h]uman beings, like most other animals, have desires and motives and are able to make choices that frequently satisfy these “first-order desires.” Human beings also have the capacity to form “second-order desires,” or preferences among their first-order desires. The characteristic that distinguishes persons from nonpersons is that persons frequently are able to make their second-order desires the basis upon which they wish to be moved to action. In Harry Frankfurt’s terminology, the essential attribute of personhood is the presence of these “second-order volitions,” which occur when the individual “wants . . . certain desire[s] to be his will.” Under this view, all that is required for moral responsibility is the ability to “act freely,” or the ability to act according to one’s second-order volition. (Farahany 2016, 66)

Notably, in this quotation, Farahany uses (as she herself explains) a language consonant with Frankfurt’s (1971, 11) vocabulary, which also refers to first-order and second-order desires.⁵ However, Farahany does not consider or comment on the fact that Frankfurt (1971 n5, 11) does not exclude the possibility that there could be “[c]reatures with second-order desires but no second-order volitions [that] differ significantly from brute animals, and, [that] for some purposes, it would be desirable to regard . . . as persons.” Although Frankfurt does not describe or discuss these creatures in this quotation, it is possible he was referring to creatures that show a high degree of intentionality, such as primates.

Similarly, Deane-Drummond (2021) has claimed that in some instances, it might be possible to also use the concept of person for animals—although this theologian reached this conclusion using a very different approach than Frankfurt. For these reasons, it is important to reflect on the notions of freedom and free will in relation to the results obtained in some experiments in which BMIs were used in monkeys.

A careful analysis of the scientific literature (which I will now briefly summarize) demonstrates that the findings in the experiments of BMIs applied to human beings and monkeys are, in fact, remarkably similar. As we shall see, the distinction between freedom of choice and freedom of action seems to apply not only to human beings but also to monkeys, which, as we know, are animals that are close to human beings on the evolutionary ladder (and thus significantly different “from brute animals,” using Frankfurt’s expression mentioned earlier).

Similarly to the cases in which BMIs were used in paraplegic patients, researchers have been able to decode the movement intentions of monkeys by following two steps: (i) first, by decoding the neural processes responsible for specific movements of the monkey’s limbs and then (ii) by using this information to build a BMI that associates the electrical impulses to specific movements of the limbs; in this way, the monkey was able to manipulate a robotic arm. Recently, some researchers have developed a new BMI technology that is able to decode the electroencephalogram signals from a monkey’s brain and transmit some electrical signals to a device, thus enabling a monkey to control the movement of a disembodied robotic arm with its thoughts. Notably, in this experiment, the robotic arm moved by the monkey was disconnected and placed at a distance from the body of the animal (Hale 2023).⁶ Thus, in this instance, the movement of the robotic arm took place without the mediation of any physical organ of the monkey, since it appears the robotic arm (detached from the body of the animal) was moved by the “thought” of the monkey. In another experiment, monkeys were able to move avatar arms that performed specific movements in virtual reality through their thoughts (Ifft et al. 2013). Arguably, the implication arising from these experiments is that monkeys are capable of performing disembodied operations driven by their thoughts.

It cannot be denied that the movement of an artificial arm from a spatial position (A) (with coordinates x_a, y_a, z_a) to another spatial position (B) (with coordinates x_b, y_b, z_b) can be described with a mathematical function. In fact, researchers who build BMIs need to use specific algorithms (utilizing mathematical-logical language) that are able to decode brain activities and transform them into electrical impulses that move the robotic arm. Nor can it be denied that such a mathematical function (like any other mathematical object) has some obvious connotations of immateriality (Anglin and Lambek 1995).⁷ Arguably, the mathematical function (with its immaterial connotations) that describes the movement of a robotic arm from position A to position B must be somehow present, perceived, and enacted in the mind of the monkey so that this animal is able to move intentionally the disembodied artificial arm from the spatial position A to the spatial position B. The enactment of a specific movement (i.e., of a disembodied artificial arm or of an avatar arm in virtual reality) must take place through some kind of immaterial faculty within the mind of the monkey, as we have seen through a BMI that uses an algorithm with a mathematical logic language. It is difficult to disprove that in these experiments with BMIs, a monkey would lack the intentionality, desire, and will associated with a specific spatial movement of a robotic arm disconnected from its body (from position A to position B), given that this movement is actually performed by this animal without the mediation of any physical organs.

Thomas Aquinas, following Aristotle, famously taught that what makes the human soul a spiritual soul, in contradistinction from nonhuman souls, is that the human soul displays spiritual faculties that do not use bodily organs (Aquinas 2008, 186 [Quaestio Disputata de Anima, a. 1, sed contra]; Aquinas, Summa Theologica, prima pars, q. 77, a. 5). In particular, Aquinas affirms “that some operations of the soul are performed without a corporeal organ, as understanding [intellectus] and will. Hence the powers of these operations are in the soul as their subject” (Aquinas, Summa Theologica, prima pars, q. 77, a. 5). The teachings of Aquinas help distinguish between mental states and spiritual states. Arguably, monkeys exhibit mental states driven by primary desires (i.e., the desire to eat), whereas human beings have, in addition to mental states (driven by primary desires), second-order desires such as the will and ability to apprehend concepts (i.e., we can understand the notion of nourishment rather than simply desiring to eat, or we can apprehend mathematical truths), or we can reflect on the self. However, as we have seen, the experiments on BMIs have shown that (similarly to human beings) monkeys can also move a disembodied artificial arm (with their thoughts and without using bodily organs). Arguably, the operation of the will (rather than the operation of the intellect) can easily be recognized in the BMI experiments on monkeys. I concur with Aquinas’s teachings that it is very unlikely a nonhuman being (such as a monkey) would have the same understanding (intellectus) as a human being in relation to things perceived in the external world. However, the results of the BMI experiments described here suggest we should reconsider our understanding of the notion of will in relation to (i) mental states (present in both human and nonhuman beings) and (ii) spiritual states (present uniquely in human beings). It seems that not only spiritual states (which are uniquely human) but also mental states (detected in nonhuman beings) show there are phenomena that reveal the existence of immateriality, given that monkeys can also move a disembodied arm with only their thoughts.

In other words, despite the fact that a monkey does not have the same understanding as a human being in relation to the ability to form a concept (i.e., the concept of what a robotic arm represents or the notion of nourishment), the experiment described shows that the movement of a robotic arm (detached from the body of the monkey) is enacted through some immaterial property (i.e., the mathematical function that enables the translation of the “thought” of the monkey through an algorithm so that the robotic arm can be moved).

Moreover, it should be noted that the monkeys, similarly to those paralyzed human patients described, needed some training before they learned to actualize their intended movements through the artificial arms. Thus, as with humans, the monkeys needed to learn to convert the information from their thoughts—conveyed through specific brain states—in order to perform specific movements of the artificial arm. In a study, it took up to fifteen days for the monkeys to learn to do this (Ifft et al. 2013).⁸ This suggests that monkeys use a similar process to humans in order to learn to actualize their desired intentions. Thus, it is very likely that the explanation provided by Farahany (described in the previous section) for the use of BMIs in human beings—in order to account for the mechanism by which the paraplegic patient deliberates and executes their actions with the artificial arm through the freedom of action— also applies to the experiments with BMIs on monkeys.

Farahany (2016, 65–66) demonstrated that freedom of action is associated with the “legal and moral responsibility” of human beings. Given the similarities between the experiments with monkeys and humans (described previously), my proposal is that her argument can now be extrapolated and applied to nonhuman beings. The fact that both monkeys and human beings seem to require some time in order to learn how to actualize the decisions that have been deliberated strongly suggests that it is reasonable to (i) distinguish between freedom of choice and freedom of action, and (ii) correlate freedom of action with moral responsibility, not only in human beings but also in monkeys. Arguably, what I have presented seems to suggest that morality is not a uniquely human phenomena but also applies (although, as we shall see, to a much lesser extent) to nonhuman creatures.

When we consider Deane-Drummond’s reflections on the notion of payback and fairness in primates in relation to morality, we could easily imagine a monkey who decides to move its artificial arm in order to get some payback/revenge on another monkey who stole its food. I argue that this possible experimental scenario provides evidence that monkeys also possess at least some level of moral agency (morality in an embryonic stage) and that their choices are actualized through their freedom of action (which is very similar to that which occurs when a human being makes a decision to act).

A human being can of course make moral decisions that have much more powerful and tragic effects on the rest of humanity and the ecosystem. For example, a man could decide (after deliberating with his freedom of action) to use his artificial arm to detonate a nuclear bomb with the desire to destroy a city, killing thousands of people. Obviously, the monkey’s notion of morality is much less developed than human morality. In other words, a monkey displays traces of moral behavior (when compared to human beings), as exemplified by Deane-Drummond’s notion of payback and revenge. Nevertheless, it should now be clear that morality is not a uniquely human phenomenon, but it is present, although to a much lesser extent, in nonhuman creatures. Thus, arguably, the results from the experiments on BMIs performed on monkeys (presented earlier) suggest that what distinguishes human and nonhuman beings is a difference in degree in the ability to make moral choices. The findings from these experiments—when read in light of Farahany’s reflections on free will in relation to moral responsibility (which I extrapolated from the human to the nonhuman world)—suggest that traces of morality can also be found in the natural inclinations of prerational beings. These inclinations are manifest through mental states that can be detected both in human and nonhuman beings. Notably, this claim does not contradict the uniqueness of human beings, given our spiritual nature and our unique ability to apprehend concepts and reflect on the self.

Moreover, my claim provides additional support to the theological approaches adopted by Porter, Hefner, Clough, and Deane-Drummond (described previously), which recognize the relation of continuity between nonhuman and human nature while maintaining the uniqueness and the special role of human beings (created in the image of God) in creation. My claim also resonates with the theological perspective of Bethany Sollereder (2016, 276) when she affirms that: “The deep commonalities between human and non-human animals do not eliminate human uniqueness, but they may increasingly turn our definition of uniqueness to one of role rather than of capacity.”

In the final section of this article, I reflect on the implications deriving from the perspective I have presented so far, specifically in relation to the fact that human beings are unique in creation since they are the only living beings created in the image of God.

Implications for the Theological Understanding of the Notion of Imago Dei

In the previous sections, I have argued that it is incorrect to insist on the endorsement of ethical theories that absolutize human reason. Instead, I propose that it is more correct to affirm that there is an ascending intensity in the phenomenon of morality as we move to higher and more complex forms of life; there is no doubt that humans have a greater level of complexity in the way in which we make decisions in relation to freedom of choice and action. Clearly, this perspective—supported (as we have seen) by the scientific observations of experimental studies of BMIs—has some repercussions for the way in which theologians should describe the notion of imago Dei. What notion of imago Dei should we then adopt, given what we know from the BMI experiments?

I believe there are two theologians—namely, Daniel Horan and Joshua Moritz—who could help in answering this question.

The Franciscan theologian Horan (2019, 560) desires “to reframe the imago Dei to include nonhuman creatures alongside members of the human species in some form.” He achieves this goal by “deconstructing anthropocentric privilege” (Horan 2019, 560),⁹ reflecting in particular on the notion of nonhuman agency. He argues that “cognition, moral reasoning, and emotion” (Horan 2019, 566) are found not only in human beings but also in other creatures. His theological approach aims to outline a vision of humanity not set apart from the rest of creation; thus, Horan aims to provide a description of imago Dei that does not create a division between human and nonhuman nature. The views presented in this article are consonant with Horan’s perspective and help in showing that there is a relation of continuity, rather than discontinuity, between human and nonhuman agency.

The second thinker is the theologian Joshua M. Moritz (2011, 307), who proposes that “[i]nstead of grounding the image of God in human uniqueness . . . [we should affirm that] imago Dei is—exegetically, theologically, and scientifically—best understood in light of the Hebrew theological framework of historical election.” A passage in his article summarizes well his position:

As human beings are the animal species historically elected, called, and commissioned by God to be His royal representatives and priestly servants who strive to accomplish God’s will for the world, human specialness lies not in the content of our characteristics, but in the very fact that Homo sapiens are the animal species who are both called and chosen as God’s image. (Moritz 2011, 329–30)

In other words, human and nonhuman beings are similar in the sense that both are created (and thus we can affirm there is a relationship of continuity between human and nonhuman nature). However, human beings reflect God’s image because they are the only creatures chosen by God to fulfill God’s plan while remaining moral agents par excellence, given their superior intellectual abilities (when compared to other living creatures). In particular, Moritz (2011, 324) refers to Genesis 1:28, 12:3, 17:2, 6, and 8 to argue for the divine election of human beings, accounting for their creation in the image of God. Moreover, he affirms: “Adam and Eve, as the primal human pair, are chosen and called to be a species of priests to the non-Adamic humans (the other hominids) and to other non-human animals” (Moritz 2011, 324). Moritz (2011, 330) further expounds this point by explaining that:

[u]nderstanding the divine likeness as election is consonant with evolutionary biology’s conception of human–animal continuity and the genealogical nature of species, and also acknowledges nonhuman animals and hominids as the ontological equivalents of humans and as fellow creatures that are substantively the same. At the same time, though, the imago Dei as election upholds the attestation of scripture and tradition in exclusively designating the image to human beings alone.

In conclusion, in this article, I have considered the philosophical-theological implications deriving from the use of BMIs in human and nonhuman beings. I have argued that freedom of action (which Farahany has demonstrated accounts for moral responsibility in human beings) is most likely present not only in humans but also in an “embryonic stage” in nonhuman creatures (e.g., monkeys.) Thus, it seems morality is not a uniquely human phenomenon but is present, although to a much lesser extent, in nonhuman creatures. As a result, when comparing human and nonhuman beings, it seems reasonable to refer to the existence of a difference in degree of ability to make moral choices. I have argued that this perspective is still consonant with some theological approaches that describe imago Dei.