Robert Kane’s libertarian theory of freedom is one of the most influential varieties of libertarianism today. Kane argues that the free will debate has been too narrowly focused on freedom of action. In his view, the fundamental issue concerns freedom of the will and how one comes to have the will one does. For Kane, the basic question is do we freely create our will? Kane argues we do freely shape our wills through a unique kind of action involving an undetermined choice, which Kane calls “self-forming actions” or SFAs. There has been much debate over SFAs and whether SFAs can relieve worries posed by familiar objections to libertarian theories. A common objection raised against Kane’s view (See Murday 2017: Mele 1999: Haji 2000: Levy 2005) is the luck objection. Kane’s critics contend that the luck involved in undetermined choices undermines our freedom and responsibility instead of enhancing them. The intuition behind the luck objection is that if an agent has a categorical power to do A or to do B with the same past and deliberation, it appears the choice to do A (or B) is a matter of luck. The problem the luck objection focuses on is the lack of contrastive explanations for choices made in SFAs. If the agent’s choice of A over B is undetermined, it seems to follow that we cannot contrastively explain why the agent chose as she did; there appears to be a missing element that is the deciding factor for her choice. In the absence of such a factor it seems to be a matter of luck as to why the agent chose as she did. This undermines the agent’s control and diminishes her freedom and responsibility. I argue that the luck objection fails to establish that the lack of a contrastive explanation makes the choices made in SFAs a matter of luck. I will do so by demonstrating that there is a non-contrastive explanation that diminishes worries about luck. I will give a brief sketch of Kane’s theory followed by a careful overview of the luck objection. I will focus on Alfred Mele’s formulation of the luck objection since it is an excellent representation of the objection and of Kane’s critics.

My discussion is divided into six sections. In section 1 I sketch out Kane’s theory and focus on his conception of indeterminism and his defense against the luck objection. Kane uses a type of control he calls plural voluntary control in conjunction with satisficing reasons to combat the luck objection. In section 2 I develop Alfred Mele’s articulation of the luck objection and explain why Kane’s reply to the luck objection is insufficient for Mele’s formulation of the objection. Mele’s argument requires providing a more robust explanation than the one provided by satisficing reasons and plural voluntary control. In section 3 I draw out overlooked dimensions of Kane’s theory and begin to develop a framework for formulating non-contrastive explanations for choices in SFAs. In Section 4 I draw out the narrative self in Kane’s view. I believe that the lack of contrastive explanations for SFAs flows from the narrative conception of the self in Kane’s theory. Through adopting the narrative view of the self we can see it is not just lucky that the agent chose as they did. In section 5 I consider how Mele might respond to my proposal. I argue that Mele’s counterargument is addressed through considering Kane’s conception of indeterminism. I conclude in section 6 that if we are self-forming beings, we must be mindful that we are still in the process of writing our life-stories and creating ourselves.

1 Self-Formation

Kane’s theory is succinctly captured by his example of a businesswoman on her way to an important meeting. As she is walking to her meeting she witnesses an assault taking place in an ally. This stirs up ambivalence in her will making her of ‘two minds’ about what she should do. She can either intervene and help the assault victim or continue on to her business meeting. She deeply desires both actions, but she can’t do both. If either action is to be chosen, it will require an effort of will to overcome the temptation to choose the competing action. She is torn between competing values, to perform the moral action, one the one hand, or to perform the selfish action, on the other. The businesswoman is engaged in parallel teleological or goal-driven processes whereby she tries to achieve multiple goals. The ambivalence in her will creates the indeterminism required to make an undetermined free choice. Kane claims we should understand this indeterminism as, “an ingredient in larger goal-directed or teleological processes or activities of the agent, in which the indeterminism functions as a hindrance or obstacle to the attainment of the goal” (Kane 2016). Indeterminism acts as a hindrance because it is an obstacle one must overcome in order to achieve one’s goals or purposes. The businesswoman must overcome the temptation to go to her meeting to help the victim, and she must overcome the temptation to help the victim to get to her meeting on time. Indeterminism understood as part of a teleological process is important for understanding how it does not undermine responsibility. The choice made as a result of this process is not random but the outcome of overcoming an obstacle to making the choice the agent was trying to make.Footnote 1

According to Kane, although the businesswoman’s choice is undetermined, this doesn’t imply that she lacks control over her choice. Kane claims she has plural voluntary control (PVC) which he uses to combat the luck objection. Kane claims agents have PVC when they can choose either competing option about which they may be deliberating and do so voluntarily (without compulsion or coercion), intentionally (on purpose) and for the reasons they want to make that choice, either way they choose (1996: 143). To frame PVC in terms of the businesswoman case, she can choose to help or to go on to her meeting. She can choose to help because she wants to be a good person, or she can choose to go on to her meeting because she chooses ambition over morality. She will also choose one option on purpose and not by coercion or compulsion.

Plural voluntary control gives the businesswoman the ability to act for either set of reasons and allows her to say she chose on purpose. PVC is part of the explanation for the choices made in SFAs. Since the businesswoman can act for either set of reasons there are satisficing reasons for either choice. Satisficing reasons are reasons that are not conclusive or decisive reasons in a sense that would make any alternative choice in the circumstances irrational or unreasonable. But they are ‘good enough’ reasons to make the choice they support a rational and reasonable one. The outcome of SFAs will be under the agent’s control and not a matter of chance because the agent will have brought about the choice made, voluntarily, intentionally and rationally, either way she chooses. Despite the indeterminism, there is a causal connection between the agent’s choice and her reasons, construed as inputs to deliberation. Kane argues that if we understand an agent’s choices in terms of this indeterministic causal framework the force of the luck objection is undermined. The businesswoman can perform either action rationally and the outcome will be causally related to the causal inputs. The causal relationship constrains agents from making any outcome come about while the outcome that does obtain will be causally connected to the agent’s will. It would be incorrect, according to Kane, to say the businesswoman’s choice is a matter of luck because the objection does not consider this relationship between the agent’s choice and her reasons and PVC. Despite his appeal to plural voluntary control, Kane’s critics maintain that the luck objection is still a problem for his view.

2 Kane and Luck

In his article, “Kane, Luck and the Significance of Free Will,” Alfred Mele argues that Kane’s view is still susceptible to the luck objection, and he appears to speak for many of Kane’s critics. Mele motivates his objection in his discussion of indeterminism. He writes,

Libertarians rightly regard determinism as a bar to agents’ having more than one causally open future (i.e., more than one future that is consistent with their past and the laws of nature), … Although indeterminism removes this bar, it also raises a worry about luck. If there are causally undetermined or indeterminate aspects of a process that terminate in, for example, a choice-- including, as in Kane’s picture, aspects that are present at the very time the choice is made and directly relevant to the process’s outcome--then, to the extent that the agent is not in control of these aspects, luck enters the picture in a way that seemingly threatens both moral responsibility and any desirable species of free agency. (97)

Mele argues that Kane’s account of free will succumbs to the luck objection because agents lack sufficient control over their choices in SFAs. Mele uses the following thought experiment of Ann to illustrate his point. Ann has made a promise to call her colleague Beth at eight in the morning to prepare her for an important meeting. Beth has recently broken a promise to Ann who is still quite furious about it. Ann believes that she ought to call Beth at 8:00 am but she also wants revenge for Beth’s broken promise. Ann makes an indeterminate effort to call Beth on time but she fails to call Beth for at least several minutes (Mele 1999). Now consider Ann* in a nearby possible world with the same laws of nature and a very similar past to Ann. Ann* is in the same conflict as Ann regarding whether or not to call Beth on time. Ann* overcomes her temptation to not call Beth on time. Are Ann and Ann* ultimately responsible for their actions? Mele thinks they are not because their choices were a matter of luck and this undermines their freedom and responsibility.

Plural voluntary control will not answer the luck objection for Mele because without a contrastive explanation for Ann’s choice, there is something that is not under her control which tips the scales and explains why she chose differently than Ann*. PVC will allow us to say both agents made rational and intelligible choices and hence their choices are not completely random. Mele grants this point but is concerned about our inability to contrastively explain Ann’s choice by appealing to features of Ann herself. Without an explanation that appeals to factors such as Ann’s character, motives, or circumstances, Mele claims that her choice is simply unlucky (Ann doesn’t call on time while Ann* does). It is not possible to contrastively explain Ann* succeeding and Ann failing because one of them made a stronger effort of will. The reason such an explanation is unavailable is that, according to Kane, such efforts are not comparable. This follows from Kane’s claim that if efforts are indeterminate we cannot compare the effort of one agent with another. If we accept this claim, then, as Mele summarizes, Ann’s choice seems to be a matter of luck because,

… given that the difference in outcome in the two cases — successful resistance and a subjectively morally proper choice in one and unsuccessful resistance and a subjectively morally improper choice in the other — is not to be explained by a difference in the amount of effort or in the intelligence of the effort, this alleged implication of the efforts' being indeterminate seems insignificant. The unsuccessful agent has worse luck than her successful counterpart. And if it were not for Ann's having worse luck than Ann*, Ann would have been in Ann*'s shoes: she would successfully have resisted temptation… (99)

Mele will not be satisfied with PVC because he wants a contrastive explanation,Footnote 2 which would identify additional features of the agent and explain why she chose as she did. What is it about Ann that prevented her from overcoming the temptation to call Beth on time? Is it her character, state of mind, or circumstances? These sorts of considerations would highlight an agential feature of Ann that could contrastively explain why Ann chose as she did. Kane’s answer to the luck objection (that PVC wards off worries about luck) will not suffice for Mele’s articulation of the luck objection.

While I agree there is no contrastive explanation for undetermined choices, I disagree that the lack of such explanations establishes that choices made in SFAs are a matter of luck. I will show in the next section that Kane can explain choices made in SFAs without a contrastive explanation and that this explanation diminishes concerns regarding the role luck plays in SFAs.

3 Value Experiments and Satisficing Reasons

In this section, I argue that while the luck objection seems compelling it overlooks features of Kane’s theory which, when brought into focus, considerably mitigate the force of the luck objection. I believe that appealing – as Kane does – to satisficing reasons alone is too weak of a basis to defend against the luck objection. Satisficing reasons allow us to say the agent acted rationally but we need a richer understanding of an agent’s decision to quell the luck objection.Footnote 3 I believe a more robust explanation for choices in SFAs is available by appealing to teleological intelligibility or continuity and the narrative view of the self. The aspects of Kane’s view I will emphasize are value experiments, teleological intelligibility and liberum arbitrium voluntatis (translated from Latin as “free judgement of the will”). My strategy in this section is to show that despite the lack of a contrastive explanation the choices made in SFAs do have an explanation. My approach will partially satisfy Mele as the aspects of Kane’s view I emphasize are agential (in the way Mele demands). It will only partially satisfy Mele because these additional features of the agent will not provide a contrastive explanation. I will argue, however, that providing a contrastive explanation in addition to the features I have mentioned is not necessary to mitigate the luck objection. My approach goes beyond the rationality involved in the indeterministic causal framework Kane deploys and focuses on the meaning (articulated in terms of values) choices in SFAs possess for the agent.

I begin by framing the discussion in terms of the indeterministic causal relationship at play in SFAs. Recall from earlier the role indeterminism plays, which is an ingredient in a larger teleological or goal-oriented process (Kane 2016, 2009). Also recall that the businesswoman is torn between competing visions of who she wants to be or to become. She has plural voluntary control and satisficing reasons that rationally explain either choice. In order to provide an explanation that is not contrastive for choices made in SFAs we also need to consider Kane’s conception of value experiments, teleological intelligibility and liberum arbitrium voluntatis. I will begin with value experiments and teleological intelligibility, which, as we will see, combine with liberum arbitrium voluntatis to explain choices in SFAs.

Value experiments (which are also plans of action or ways of life) are initiated by agents in every one of their self-forming actions (Kane 1996). SFAs are the result of a conflict in the will but this conflict is not a struggle between mere desires but between embedded values within the agent’s will.Footnote 4 The agent is deciding which set of values they want to try out. Their past influences but does not determine their choice so that there are multiple causal inputs available for the agent to choose. None of them are decisive inputs but are inputs which are meaningful. The choices made in SFAs require understanding the choice in terms of meaning for the agent and the value experiment they have chosen. In deciding what to do the businesswoman is effectively saying, “Let’s try this. It is not required by my past, but it is consistent with my past and is one branching pathway my life could now meaningfully take. I’m willing to take responsibility for it one way or the other…. I will be aware that it was my doing and my own self-making either way.” (Kane 1996: 145).

Kane discusses an important element of value experiments that is illuminating for how we non-contrastively explain free choices beyond appealing to satisficing reasonsFootnote 5 (Kane 2010). This is the element of risk involved in value experiments which follows from being unable to know the result of a value experiment ahead of time. It is a risk for which the agent will take responsibility. This leads into the narrative explanation that goes beyond satisficing reasons. Whichever way an agent chooses, the narrative explanation highlights the risk involved in experimenting with the way of life she chooses. As Kane points out, “What makes value experiments ‘experimental’ is … that the commitments with which we get into them are no guarantee of how they will turn out.” (Kane 2010: 82). This means that while satisficing reasons rationally explain the agent’s decision, the explanation I am proposing takes note of the nature of the value experiment the agent initiates. For the ultimate success of one’s value experiments is, “satisfaction with one’s life as a whole” (Kane 2010: 85) (emphasis mine). Value experiments need to be placed within the context of teleological intelligibility as the judgement of their success is being satisfied with one’s narrative or life story. I will give an account of what this kind of satisfaction is grounded on in the next section.

Teleological intelligibility requires free choices to have a “narrative continuity” or to fit into “meaningful sequences” (Kane 1996: 146). Teleological intelligibility or continuity does not require the agent’s past be the sole basis of explaining the agent’s free choices, but it frames their decision. When the agent decides what to do, their decision must be consistent with their past.Footnote 6 This is not to say that there is only one choice agents can make but rather that there are multiple meaningful choices which are influenced by the past but are not required by it. Teleological intelligibility highlights the meaning of the choice for the agent as the options are meaningfulFootnote 7 to her. The fact that the options were meaningful is the source of the conflict in her will and what made it ambivalent in the first place. As an agent’s will is ambivalent in SFAs, the structure of her will (i.e., her values, preferences, desires) is being reshaped as well. The narrative framework of her decision plays a role in framing how she will reorganize her will.

We can illustrate how, in other examples of self-formation, such as a prudential choice, we would not say it was luck that caused the choice. For example, I want to be healthier and live longer, but I also want to drink more wine and read philosophy. It is important to keep this in mind because it further brings out how, in other instances of self-formation, we would not say the choice was a matter of luck. The me in the actual world could decide to drink more wine while the counterfactual me could decide to exercise instead. Both agents are deciding to pursue meaning they believe is worthy of pursuit. It is not just a matter of luck which choice I make because we can explain with reference to me (the agent) why I chose as I did. The narrative framework in which SFAs occur provide a non-contrastive explanation for all self-forming choices. Through focusing on the conflicting values that create the ambivalence in agents’ wills we can see how narrative explanation non-contrastively operates. The businesswoman is making a moral choice and her decision is understood in the same teleological terms as my prudential decision. Plural voluntary control can give both of us the ability to choose rationally but a richer, more robust explanation for the choice is cast in terms of the meaning it will give each of our life stories.

Explaining the choice in terms of the meaning it gives one’s life story improves on the explanation offered by satisficing reasons. The explanation is still not a contrastive explanation but nonetheless offers an explanation that makes choices less susceptible to worries about luck. It builds on the explanation of satisficing reasons by adding narrative features of the agent to the explanation. The values that are being tested are agential properties as they are embedded within the agent’s will. Value experiments are not an external object that the agent could appeal to in order to explain what they did. They are an internal matter involving different aspects of themselves. The worry about luck in Mele’s case arises from there being no contrastive explanation that appeals to a feature of the agent (character, motive, etc.). If, however, we explain Ann’s choice in terms of the meaning not calling Beth on time gives her life story, our understanding is not inhibited by the lack of a contrastive explanation. Ann’s decision is understood as resolving a conflict in her will between embedded values. Her choice is risky because she is not decisive in pursuing one set of values over the other. This explanation is distinct from the one given by satisficing reasons because satisficing reasons only demonstrate the rationality of either choice that is made. They do not emphasize the values in conflict within the agents will, which my account does. To fully understand how the account I am arguing for overcomes the luck objection, we need to flesh out liberum arbitrium voluntatis and connect it to the points made so far.

We can see how liberum arbitrium voluntatis is used in the narrative framework through an analogy of Kane’s (Kane 1996: 146–147). Suppose a writer is in the middle of writing a novel. The heroine of the story faces a conflict like the businesswoman and, just like the businesswoman, the heroine’s character is not sufficiently developed for the author to determine what she will do next; there is no decisive path for the heroine to take (ibid). Yet, the past provides a limited range of inputs in terms of satisficing reasons to guide the author’s decision. Agents in SFAs are analogous to the novelist in instructive ways. For if agents are truly free as Kane requires, they are making themselves from a past which does not limit them to one future pathway. As Kane writes,

If these reasons were not sufficient or conclusive reasons, that’s because, like the heroine of the novel, I was not a fully formed person before I chose (and still am not, for that matter). Like the author of the novel, I am in the process of writing an unfinished story and forming an unfinished character who, in my case, is myself. (2009:43)

Being the author of, and character in, one’s own life story is what value experiments, teleological intelligibility and liberum arbitrium voluntatis are meant to capture. For if we are self-forming beings, we are guided by our past but not determined by it as the future is meaningfully open to us. When we are in a moment of self-formation, our decision is, as I’ve argued, to pursue a way of life we feel is worthy of being pursued. We might come to change our minds about the worth of a way of life but we will take responsibility for it if we freely made the choice. When we combine these elements of Kane’s view they undermine or at least mitigate the luck objection. I think this conclusion is supported by the narrative framework because it provides an explanation of undetermined choices that, while not contrastive, makes the agent play an important role in the explanation. To see how a non-contrastive explanation in the manner I am advancing is not vulnerable to concerns about luck let us return to Mele’s thought experiment of Ann and Ann*. Both of their choices are explainable as pursuing a set of values they believe were worthy of pursuing. Ann could plausibly be said to continue her narrative by pursuing values which believed in holding grudges while Ann* decided to try out being someone who does not hold a grudge. The lack of a contrastive explanation does not appear to me to raise any problems regarding luck and control. Both of their choices are explained in terms of what they value in addition to having plural voluntary control.

This explanation of Ann and Ann*‘s choices are like my earlier example of making a prudential choice. The explanation makes the agent an important factor in the explanans and while the explanation is not contrastive we can see that the choice was not just a matter of luck either. The agent’s values and narrative structure are part of the agents themselves. Hence, there is an explanation of the kind Mele demands that emphasizes agential features in the explanation of an undetermined choice. The narrative framework I have been urging builds on the model of explanation Kane offers with satisficing reasons and improves on it. In addition to the rationality of agents’ decisions offered by satisficing reasons we have an understanding of why agents chose as they did. We can understand an agent’s choice as deciding between meaningful options and taking the risk to pursue one way of life over the other. This is a risk for which the agent will take responsibility. As the authors of our life stories we freely take on this risk knowing it, “may make me a (morally) better or worse person or a happier or sadder person” (Kane 1996: 146). This is what it means to be a self-forming being.

4 Narrative Selfhood and Values

I have up to this point defended Kane’s theory on the grounds that there is a broader narrative structure in which choices in SFAs are made. I have argued that value experiments take place in a broader narrative framework. What remains to be seen is how the values become embedded within the agent’s will, given that they are being experimented with in SFAs. How embedded can they be if we are testing them out? There is also the matter of fleshing out liberum arbitrium voluntatis in narrative terms since it is a fundamental part of Kane’s reply to the luck objection. We also need to flesh out the relationship between satisfaction and narrative unity. For these reasons further exposition of the narrative view of the self is warranted. In this section, I will situate the narrative self in the literature and attempt to put forward a synthesis of narrative views of the self and Kane’s account of free will.

Recall, Kane states the problem of free will is a problem regarding self-formation and understood in these terms, narrative selfhood, as we will see, goes hand in hand with self-formation. Narrative views of the self are quite broad (See Schechtman 2011) so I will be focusing on one kind of narrative view: the hermeneutical approach. The most influential of the hermeneutical narrativists is arguably Alasdair MacIntyre and I will draw from his book After Virtue. Through using the narrative approach, I will demonstrate what it means for the self to be defined by its values and for those values to be embedded in the agent’s will. The self that emerges from this account is one that is narratively self-constituted.Footnote 8 I think this fits nicely with Kane’s claim that self-formation is a gradual process. The values that are embedded in our wills become embedded through the gradual creation of a back-log of self-formed character (Kane 2011: 396). I think the narrative self can help us to understand how agents decide to create their characters in such a gradual manner. Furthermore, it deepens my account of non-contrastive explanations for choices made in SFAs by going beyond the rationality and causal role of satisficing reasons. I will begin with an exposition of the hermeneutical narrative self.

The hermeneutical view claims selfhood is necessarily tied to agency and agency requires narrative. To be a self on this view requires teleological intelligibility, which entails casting the explanation for an agent’s actions as the end or telos he or she pursues (Schechtman 2011: 396). The main defenders of the hermeneutical account are Charles Taylor, Alasdair MacIntyre and Paul Ricoeur (Behrendt 2014: 338). MacIntyre’s narrative account argues that human action needs to be understood in narrative terms in order to be intelligible. To explain an action one must, “identify it under a type of description which enables us to see that occurrence as flowing intelligibly from a human agent’s intention, motives, passions, and purposes” (MacIntyre 1984: 209). A well-lived life on this view is one engaged in a quest for unity in one’s life. The unity of one’s life is achieved through striving for the attainment of one’s telos (Ibid). The idea that one’s life is unified through striving for one’s telos is also what attracts the most criticism from anti-narrativitists such as Galen Strawson. Strawson, and other anti-narrativists, claim it is not descriptively true all (non-pathological) people aim at one chronological end point that gives their life narrative unity and meaning (Strawson 2004).

This objection distorts MacIntyre’s account, however. For MacIntyre, the concept of telos ought to be understood as “an end or goal not in the sense of the thing aimed at but in the sense of the agent’s aiming at that end” (Annas 1993: 34). To contrast it with Strawson’s interpretation, Strawson sees telos in terms of concrete goals whereas MacIntyre should be taken to be employing telos in a broader sense. This is more in line with MacIntyre’s employment of telos as a quest with his claim, “the good life for man is the life spent in seeking for the good life for man” (emphasis mine) (MacIntyre 1984: 219). Kathy Behrendt call this the agent-process interpretation of telos (Behrendt 2014: 339). This interpretation of telos disarms Strawson’s critique because it does not conceive of telos as a definitive goal. The agent-process view conceives of telos as having multiple avenues for achieving one’s telos. It is a quest we embark on not a destination we seek. As Anthony Rudd writes in his defence of the agent-process view, “There is no sense in supposing that we could say--well I’ve attained narrative unity, now what shall I do?” (Rudd 2007a: 67). Rudd’s point is that it makes no sense to interpret telos as a definite end-point. Narrative unity is unintelligible as a concrete goal. This raises the question of what telos is if these interpretations are correct in articulating what it is not. I will elaborate what telos is in my view below through connecting telos to Kane’s articulation of a narrative self.

We can begin to understand how by bringing in Kane’s idea of will-setting. An agent sets her will in SFAs through freely choosing her action either way she chooses. The businesswoman sets her will towards being an altruistic or selfish person through her decision. Setting one’s will is part of the larger picture involved in SFAs. We are gradually setting our wills through every SFA we engage in. The businesswoman has presumably a back-log of self-formed character where she is inclined to act in certain ways through choosing the character, purposes and ends she possesses (Kane 2011). Recall, I said earlier that the values embedded in an agent’s will were gradually created. We can make sense of how an agent experiments with her embedded values if we understand that they stem from the whole of one’s life story as it has developed. We are not determined to act a certain way but our actions, on Kane’s conception of libertarianism, must be connected to one’s past. The experimentation with values is not to be understood as trying something new in the sense of being disconnected or not causally related to one’s life story. Rather, we should understand it as deciding which path in the garden of forking pathwaysFootnote 9 we want to try out.

I think will-setting and self-formation can be further understood through incorporating narrative unity as I’ve interpreted it. SFAs are specific moments in our life histories and we should understand the unity agents strive for in narrative terms; specifically, in the agent-process manner of narrative unity. We are not aiming at a chronological end point but we are trying to see if a specific way of life will guide to us to living a satisfying life. The relation between unity and satisfaction is understood as attaining unity in pursuit of satisfaction (which I will discuss in a moment).

For example, imagine a philosophy studentFootnote 10engaged in an SFA over which graduate school she should attend. She is not thinking about what her life story will be 50 years from now but where she wants to go for the near-future. Her choice will have caused her narrative to go in the direction she chooses. While a contrastive explanation is not possible, we will understand why she made her choice because of the way the choice is embedded in her narrative framework. Rudd (2007b: 543) accurately describes this as follows: “Having a narrative sense of someone’s life … is indeed presupposed by the use of narrative to explain particular actions… To understand an action is to understand why an agent performed that action, and that involves understanding something about the agent in question.” The agential properties required for explaining a choice made in SFAs are value experiments, teleological intelligibility and liberum arbitrium voluntatis. They are as I argued internal aspects of an agent’s will or identity. An SFA’s teleological intelligibility is fundamentally grounded in narrative terms because of the causal and meaningful connection between the past and the meaning the choice will give to the present and future. As Rudd writes,

To think of who I am in narrative terms is to think of myself as a temporal being and so as a being with a past; but also to understand my current situation as the result of my past history, not just causally, but in the sense that my past history has established the meaning of my present situation. Moreover, it involves thinking of myself as an agent… and therefore as someone who is partly responsible for the shape his life has taken (Ibid).

This is how agents are deciding when engaged in an SFA in my view. The meaning of each choice in an SFA has the significance it does because of its connection to the past. To achieve narrative unity and set our wills in SFAs is to take the risk to try one way of life. Recall that the success of value experiments is satisfaction with one’s life as a whole. We do not know ahead of time if our attempts will be successful but, if we follow Kane, it is the risk to do so that is what we freely chose and through choosing gives us temporary unity. As Rudd writes, “For most people, what unifies their lives is not one project among others which is given overriding importance, but rather, an attempt to lead a fulfilling, satisfying life.” (Ibid 545). The unity we achieve is not a destination we aim at but a pathway or method of structuring our wills (Behrendt 2014: 340). As a method, narrative unity is understood as reconciling different parts of who we are and experimenting with different ways of life. The ultimate justification for a way of life lies in the future so we strive to try and see if it really does lead to a fulfilling life.

If the test of value experiments is satisfaction, then what does it mean to be satisfied with one’s life story? Kane (2010) argues that satisfaction with one’s life as whole is understood in terms of happiness or flourishing as the ancient philosophers understood happiness:

The ultimate test of success for value experiments of individuals … is happiness or flourishing understood in the sense that the ancient philosophers described as ‘satisfaction with one’s life as a whole’… to understand what happiness is, one would have to consider how the many purposes and interests of life fit together in a meaningful way and what purposes and interests are ultimately worth pursing (85).

Kane (1996) also argues that the test for value experiments is happiness but what happiness is not a concrete goal. This follows from Kane’s claim that what is radical about value experiments is that it is not merely about how one attains happiness but about “what happiness is” (Kane 1996: 209).Footnote 11 The value experimentation inherent in free will implies that the rightness or wrongness of our free choices cannot be known with certainty in advance (a priori) (Ibid). We can connect this idea with MacIntyre’s earlier claim that the good life is in search of the good life in the form of a quest. This is evident in what Kane calls “value empiricism”Footnote 12 which is opposed to value rationalism. Value rationalism, according to Kane, is the view that we can answer questions about fundamental values by constructing necessary premises that no rational person could reject. Value empiricism rejects this and claims, “we address ultimate questions of this kind through freely chosen experiments in living and cannot settle them by a priori reasoning from certain or necessary premises that no rational person could deny.” (Kane 1996: 209) Value empiricism does not imply value relativism as agents on Kane’s view must believe certain values are more worthy of pursuit than others.Footnote 13 If this were not the case, agents would never struggle or be ambivalent about what to do. The businesswoman would simply pursue her self-interest and continue on to her meeting without hesitation (Ibid). As Kane writes, “If one had such certainty, there would be no incommensurability and plural rationality prior to self-forming choice. What can be established prior to self-forming willings is that a plan of action or way of life is worth trying as a value experiment (that, indeed, is the goal of deliberation), but not that it is the right choice with certainty in advance--not if we are engaged in ultimate self-formation...”Footnote 14 (Ibid).

I think this leads to understanding the narrow arbitrariness permitted in SFAs as an implication of having an unfinished character. We are unfinished selves because self-formation is a gradual process and a matter of degree. It is a cumulative process where we build our self slowly in our attempt to pursue a fulfilling and satisfying life. Since we are not finished creating ourselves there is no decisive path for us to take. Thus, freedom, for Kane, is in a sense tied to our personal identity. Kane’s approach to the free will problem has been to argue that free will should be thought of in terms of what we are or who we are (Kane 2009). Self-formation is arbitrary in the permissible sense of liberum arbitrium voluntatis because who we are is (to a degree) ultimately up to us. It is not arbitrary in a broader sense because the narrative framework gives us an understanding of how we decide who we want to become. The relationship between an agent’s choice, her value experiment and the narrative framework it takes place in is not arbitrary in a problematic sense. On my account, an agent’s choice reflects the pursuit of a set of values embedded within her will. It reflects agents choosing to continue their quest in seeking the good life and answering fundamental questions about values. To freely choose one set of values is to take the risk to see if the way of life chosen was a good one for the agent. It is risky because agents believe some ways of life are more worthy than others even though they are strongly attracted to all of them (Kane 1996:205). The risk of failure is what defines any genuine experiment whether it is scientific experimentation or experimenting with values.

The focus of the luck objection is on how an agent can choose to pursue one way of life over another. Mele’s articulation of the objection captures the heart of the objection that undetermined choices appear arbitrary and just a matter of luck since they lack contrastive explanations. What is it about the lack of a contrastive explanation that makes undetermined choices seem arbitrary? Kane’s discussion of this intuition focuses on what he calls ‘antecedent determining control’ (Kane 1996: 144). Antecedent determining control is an agent’s ability to “be in, or bring about, conditions such that one can guarantee or determine which of a set of outcomes is going to occur before it occurs, whether the outcomes are one’s own actions, the actions of others, or events in the word generally” (Ibid). We can have contrastive explanations of our choices in cases where we exercise antecedent determining control. We exercise such choices when our actions are determined by our self-formed character or will. When Martin Luther uttered the words “Here I stand I can do no other” (Dennett 1984: 561) he exercised antecedent determining control. He acted from an already formed character which allowed him to have the control that guaranteed he could make his utterance. We can have a contrastive explanation in these cases because they can explain guaranteed outcomes through using determining factors of one’s character. In the absence of such a guarantee Mele’s objection argues that SFAs must be just a matter of luck. I think this is too hasty a conclusion and my proposal attempts to show how the absence of a contrastive explanation (and therefore the guarantee of a result) does not lead to mysterious actions or reduce SFAs to just a matter of luck.

I think the narrative framework can give us everything we need to make self-forming choices intelligible. The fact that the explanation is not contrastive does not leave us with mysterious inexplicable actions. The explanation runs deeper than simply showing that an agent’s choice is rational, intelligible and caused. The explanation I have argued for acknowledges that choices in SFAs have rational satisficing reasons, but offer much more, besides. The focus (and therefore the difference from satisficing reasons) is on risk, values and satisfaction understood as happiness. Given that the narrative framework requires knowing something about the agent (of which risk, value experiments, narrative and satisfaction are a part), and I have provided that requirement, I do not see what is necessary about contrastively explaining an agent’s choice. If the explanation were contrastive, then the choice would not be risky and therefore not a genuine experiment of one’s embedded values. The outcome of the experiment would be guaranteed beforehand and therefore not be a genuine experiment, given how experimental was earlier defined. The narrative model allows us to see why, despite the lack of a contrastive explanation, it is not just lucky that the agent chooses as she does through the focus on properties internal to her will.

5 Narrative and Luck

I have up to this point tried to answer the luck objection as Mele has formulated the problem by identifying agential features that non-contrastively explain choices in SFAs. I have argued that by placing one’s value experiment into a larger narrative structure we can understand why one acted as one did. It is not necessary to contrastively explain choices in SFAs in addition to providing the features of the agent I have outlined. Mele might make the following reply to my argument: “You can explain Ann’s choice in terms of her decision to engage in a particular value experiment, but it’s just a matter of luck that she picked this experiment rather than the other, so we are no further ahead, are we? After all, since Ann and Ann* have sufficiently similar pasts, either choice (of value experiment) will fit equally well with their overall narrative structures as it does with their satisficing reasons.”

Mele’s rebuttal claims that while I can provide an explanation of Ann’s choice in terms of her value experiment, I have only delayed the luck objection, since Ann’s counterfactual self Ann* has a similar past and with both of their choices being undetermined and fitting equally well with their overall narrative structures, the resulting choice made is just a matter of luck. Mele’s objection places a lot of weight on the indeterminism involved since without a contrastive explanation neither Ann nor Ann* can guarantee the result of her choice before she chooses. This illustrates the connection between luck, control and responsibility and what Mele and other critics are worried about. That if the choosing of one value experiment rather than the other value experiment is undetermined it is just a matter of luck which experiment is chosen. The objection appears to pose a worry about the lack of control over choosing one’s value experiment. We can begin to answer this worry through recalling my earlier discussion of indeterminism for Kane and its implications for luck, control and responsibility. In the first section while outlining Kane’s position I highlighted how Kane conceptualizes indeterminism as, “an ingredient in larger goal-directed or teleological processes or activities of the agent, in which the indeterminism functions as a hindrance or obstacle to the attainment of the goal” (Kane 2016). Kane claimed that indeterminism acts as a hinderance because it is an obstacle one must overcome in order to achieve one’s goals or purposes.

Consider two examples Kane cites for the purpose of demonstrating that indeterminism does not undermine control and responsibility. The first involves an assassin trying to kill the Prime Minister with a high-powered rifle. The second case is of an angry husband who while arguing with his wife angrily throws down his arm with the intention of breaking her favourite glass top table (Ibid). Kane supposes that in each case there is some indeterminism present in each agent’s nervous system, in their brains or arms, which is a hinderance or obstacle to their efforts that they must overcome. It is thus uncertain whether they will succeed in accomplishing their goals or purposes but if the agents do succeed in achieving their goals nonetheless, despite the indeterminism—if the assassin succeeds in killing the Prime Minister and the husband succeeds in breaking the table—they are both fully responsible for their actions because they succeeded in doing what they were trying and intending to do, in spite of the indeterminism present, and not because of it.

Kane (2007) supports this argument through imagining three assassins, each of whom killed a Prime Minister. Kane supposes that each of the assassins had a different chance of success at accomplishing their goals. One assassin had a 50% chance due to indeterminism causing her arm to waiver. The second assassin had an 80% chance and the third assassin had a nearly 100% chance of succeeding. If they all succeed, nonetheless, despite the different probabilities of succeeding, they would all be fully responsible. It would be absurd to suppose that one was only 50% responsible, the others 80%, and near 100% respectively, or that one deserved a 50-year prison sentence, while the other two deserved an 80-year sentence and a 100 years sentence. They are all equally guilty if they succeed, because they succeeded in doing what they were trying and intending to do.Footnote 15 Just because there is a chance one might fail at doing what one is trying and intending to do, this does not entail that, if you are successful in spite of this possibility, you are less responsible. Connecting this point to SFAs, indeterminism function in SFAs in the same manner it does in the case of the assassins, as “an ingredient in larger goal-directed or teleological processes or activity, in which the indeterminism functions as a hindrance or obstacle to the attainment of the goal” (Ibid 35). The processes that indeterminism is a part of are, in the case of SFAs, effort of will, with each effort having the goal of making one of the choices rather than the alternative one, and thus successfully resist the temptation to make the alternative choice. The indeterminism makes the attainment of the goal open to the risk of failure but when one of the choices is made, the goal of the respective effort will be attained, despite the indeterminism and not because of it. The choice made is not merely random or a matter of luck, rather it is the voluntary achievement of a teleological process initiated and sustained by the agent, who has succeeded in bringing about the purposes they intended and tried to bring about. They are, again, responsible for the outcome despite the indeterminism either way the agent decides to choose. This is supported by the discussion of Kane’s assassins because the assassin with 100% bears equal amounts of responsibility with the assassins with 80% and 50% chances of success, if they do succeed in attaining their goals despite the hinderance of indeterminism to their goals. This holds true even if the assassins with 80% and 50% had only a 30% and 20% chance of succeeding in attaining their goals.

This is important to consider because it is connected to Kane’s claim regarding the probability of choices in SFAs. Kane asserts that, “The initial settings of the probabilities in free choice are related to the agent’s prior character and motives and may be anywhere between 0 and 1 … depending upon the strength or weakness of the agent’s will with respect to the options… depending upon the comparative strengths of the agent’s inclinations to act… The equiprobability setting (.5,.5) is thus only one of many possible ones” (Kane 1996: 177–178). This claim regarding the probability of choices is properly understood in the context of the present discussion concerning how indeterminism does not undermine responsibility. If we keep in mind that agents make free choices not because of indeterminism but in spite of it, this goes a long way to answering concerns raised by Mele and others, that the choice of one value experiment rather than another was just a matter of luck and thus not responsible for the outcome. To put it in terms of Ann and Ann*, they are just like the three assassins who are attempting to do the same action but with different chances of success. Just like the assassins, Ann and Ann* would nonetheless be equally responsible for the value experiment they chose to initiate, if they are successful at attaining the initiation of the value experiment they were trying and intending to initiate. In the case of SFAs agents are making efforts of will to initiate their value experiments. Ann cannot be held only 60% responsible while Ann* is held 40% responsible, if they succeed at what they were trying and intending to do. That would be just as absurd as it would be in the case of the assassins where each assassin is given a prison sentence based on the probability of the success of achieving their goals and overcoming the hinderance to their goals. I think this directly answers Mele’s objection about the lack of control and responsibility over undetermined choices.

6 Conclusion

If what I have argued for is correct, then the luck objection seems to be in jeopardy. The luck objection is answered through the proposed framework by highlighting features of agents (their values and life story). My framework explicitly does this work through explaining the agent’s choice as what values the agent decides to pursue. When the neglected aspects of Kane’s view are combined as I have arranged them there is nothing which makes free choices lucky in any problematic sense. Plural voluntary control is not sufficient to combat the luck objection but put in conjunction with an agent’s narrative framework, which is the basis of the choice from the agent’s perspective, the luck objection is undermined. For plural voluntary control gives us a causal relationship where the choice is causally related to the causal input while my framework connects the explanation of the choice to the agent; her value experiments and her narrative structure. Both of these considerations demonstrate agents free choices are not just a matter of luck. Additionally, I argued that focusing on indeterminism as a hinderance to attaining one’s goals answers any worries that if SFAs are undetermined, the value experiment chosen was just a matter of luck. There is perhaps a degree of arbitress left in SFAs since the choices in SFAs are both meaningful but this concern is answered through the relationship between free choices, value experiments and narrative structure. Yet as we’ve seen, the arbitrariness is what follows from being an unfinished person whose narrative will be unified either way she chooses. In our attempts to live a satisfying whole life, it is not problematic to take the risk involved in pursuing our value experiments.

If we truly are the authors of, and characters in, our own life story which we freely create then we must be mindful we are not done writing our story. We are in pursuit of a narrative unity which we only have while our will is settled.Footnote 16 The narrative structure of our will and life story becomes unravelled as a result of the conflict that creates SFAs. Indeed, SFAs are the moments in our lives where we decide how we want to pursue our life-stories through the value experiments we initiate and the narrative framework they are initiated within. If the luck objection is to survive it would have to undermine the narrative explanation by showing the lack of a contrastive explanation still raises the worry about luck, even in a narrative context.