Ferenc Huoranszki is a Central European philosopher whose 2011 book, Freedom of the Will: A Conditional Analysis is a defense of "compatibilist free will," based on what he calls the "conditional analysis" of doing otherwise, namely, the agent could have done otherwise IF he had chosen to do otherwise. This he traces back to Hobbes, Hume, and Locke, and he cites the explicit example of G.E. Moore Huoranszki accepts physical determinism, that events can happen only as they actually do happen. He denies that "determinism at the level of fundamental physical laws is incompatible with agents having free will." (p. 3.) Like John Martin Fischer, Huoranszki regards "the issue of free will only as a question about the conditions of responsibility for our behavior." (p. 6.) He is right that "many things originate in us and are attributable to us even if we do not do them of our own free will," so that we can be held responsible even if we were not "free" to do otherwise. But this should be grounds for separating free will and moral responsibility, not conflating them. The problem of free will and determinism is a scientific question. The problem of moral responsibility and appropriate reward or punishment is a social and cultural question. For a compatibilist, Huoranszki has a very sympathetic understanding of Robert Kane's "plural rationality" and Self-Forming Actions, and he offers a strong defense for them against Richard Double's criticisms. He starts with Peter van Inwagen's idea that free decisions where the agent can do otherwise are restricted to three special occasions - Buridan's Ass cases (the liberty of indifference), when our duty is in conflict with our interests (Kane's Self-Forming Actions are here), and when we have to choose between incommensurable values. (p. 151) All these cases are what Huoranszki calls "indifference cases." But can libertarians avoid the conclusion, he asks, " that the ultimate grounds of our responsibility must be our indifference?" "What free will would require is that our preferences be undetermined because our reasons do not determine our judgment about the best course of action." (p. 159) This sounds very much like Kane's Self-Forming Actions. "Why would our responsibility be rooted in the fact that sometimes our reasons fail to determine what we judge to be best," he asks, and then explains,
In such cases, agents are not determined to act upon one set of reasons. But the chosen action is undetermined not because the agent's action might go against her reasons, but because it is undetermined whether one set of reasons rather than another will eventually cause her action. Hence, although agents' rational preferences (what they judge to be good) causally explain their actions, it is up to them which set of their reasons shall cause their action. Robert Kane calls this requirement about libertarian free will dual, or later, plural rationality (Kane 1989; Kane 1996: 107-109). Importantly, plural rationality, despite the way it is called, is not a requirement about the morally responsible agent's abilities. Rather, it is meant to refer to a particular mental state or condition. The ability that ultimately grounds agents' responsibility is volitional. But agents can exercise that capacity only in those conditions in which their volitional effort must decide which of the two or more potentially motivating sets of reasons will become causally operative in the production of their action (Kane 1997: 127).15 The idea is that volitional control requires the presence of at least two competing sets of reasons. Both sets of reasons must have some motivational potential. In some situations, however, it is undetermined which sets will actually motivate the agent. But whichever set of reasons has actually become operative in the production of the action, agents perform the chosen action for a reason. Interestingly, this libertarian account - just as certain versions of compatibilism — presupposes reasons dependence. For the actions that ultimately ground agents' responsibility must depend on their reasons. Plural rationality is a condition of freedom of will not because agents can act against their reasons, but rather because they themselves can choose, by their own volitional effort, the reasons that will cause their actions. If libertarian free will is understood in this way, then the indeterminacy of choices does not imply that agents cannot rationally control what they do. Agents' actions can be controlled by their reasons; it is the motivating force of their reasons that is controlled by their volitions. To estimate its plausibility, it is instructive to compare the plural rationality strategy with the indifference strategy. The plural rationality strategy has the obvious advantage that it does not require indifference in any sense. In order to be in the state of plural rationality, agents are not required to be equally motivated to do two or more incompatible actions. They only have to have conflicting reasons, all of which have some motivational potential. Which reason will actually become operative and thus which action will be performed is, however, up to the agent. It is the agent's own volitional effort that will decide.
Huoranszki points out that Kane's plural rationality argument does require the ability to do otherwise (the availability of alternative possibilities), nor does it even assume that the resulting decision is "free" in the sense of undetermined by reasons, character, motives, etc. He quotes Kane,
Huoranszki's claim that we cannot determine our own selves is a variation on Galen Strawson's Basic Argument that a causa sui is impossible.An adequate account of incompatibilist free will should allow that some (and potentially many) everyday acts explicable by reasons may be caused or determined by characters and motives already formed. The possibility of such actions, as I see it, is part of a complete theory of ultimate responsibility. (Kane 1996: 120)Thus, Kane accepts that agents can be responsible even if they could not have done otherwise at the time when they perform the action. We have seen already that every theory of the freedom of will as a condition of agents' responsibility must accommodate this possibility. But Kane, following Dennett and van Inwagen, does not only claim that there are special situations in which we are responsible even if we lack the ability to do otherwise. They assume that most of our actions that are explainable by our motives and character are not freely willed at all. Most motivated actions cannot be spontaneous in the sense that they cannot depend on the agent's choices and hence cannot be freely willed either. It is for this reason that they must introduce a strong and — as I shall argue in the next chapter — unsatisfiable condition of responsibility: the condition that agents must determine their own self. Kane says that when agents' actions are determined by their character and motives they are responsible to the extent that they are responsible for being the sort of person they had become by that time (Kane 1996: 39). And Peter van Inwagen claims that 'it is an old, and very plausible, philosophical idea that, by our acts, we can make ourselves into the sorts of people we eventually become' (van Inwagen 1989: 420). It seems, therefore, that both the indifference strategy and the plural rationality strategy rest on the possibility of ultimate control. According to both, agents are responsible even if their character and motives determine what they do. But agents are responsible only because their own character and motives have been self-determined by their earlier undetermined free actions. And it is exactly here where I see the fundamental difficulty with these libertarian views. According to the compatibilist view I endorse, we are responsible only if our actions are self-determined in the sense that we ourselves determine what we do or what we are able to do. But libertarian 'ultimate responsibility' implies that we are responsible only if we can determine our selves. Both libertarian views which we have considered assume that our responsibility for our actions is rooted in our responsibility for being the sort of persons we are, or that 'by our acts, we can make ourselves into the sorts of people we eventually become'. But, as I shall argue in the next chapter, we have serious reasons to doubt both that 'ultimate responsibility' exists and that it is necessary for responsible agency.
There are well-known arguments aiming to prove that free will and responsibility are impossible, no matter whether our world is deterministic or not. According to those arguments, we could be responsible only if we were causa sui in the sense that we determined our own self. For how we are, mentally speaking, determines what we do. And even if how we are, mentally speaking, is determined by our earlier choices, since those choices are also determined by how we were, mentally speaking, at an earlier time, responsibility would require that agents themselves create their own self. But no finite being, like us, is able to do that. So we are never 'truly' or 'ultimately' responsible. Some libertarians hope to answer this difficulty by claiming that there are some 'buck-stopping' undetermined actions by which agents can ultimately control how they are (Kane 1996: 114). According to them, we are responsible for our other actions only because they are the results of these self-forming and hence buck-stopping actions. In the last two chapters I argued that libertarians cannot answer the causa sui objection in this way. But I have also claimed that self-determination in the sense in which it is indeed a necessary condition of responsibility in certain contexts does not depend on our alleged capacity to determine our own self. Rather, in order to be responsible for our actions, we ourselves must be able to determine some of our future abilities and the situations in which the lack of our ability can matter for our moral responsibility. But this notion of self-determination does not require either that we determine how we ate, mentally speaking, or what reasons we have for our actions. Our reasons, character,- and motives are not abilities, neither are they factors that can deprive us of any responsibility-relevant actional abilities. Even if they may constrain to a certain extent what we do, they surely do not make us unable to perform actions that we have actually failed to perform. As we have seen in Chapter 6, sometimes we explain agents' actions by their reasons, at other times we explain what they do or fail to do with reference to their character or motives. But in whatever way we explain agents' actions, as long as they have the ability to perform an actually unperformed action, it is up to them, in the relevant sense, what they do.