Agency, Freedom and Moral Responsibility consists of sixteen independent essays — each written by different authors, and an introduction by the editors. The common ground for the essays is that the authors (with a few exceptions) participated in at least one of the two symposia (2008) “Belief, Responsibility, and Action” and (2012) “Alternatives, Belief and Action” that took place at the Faculty of Philosophy at the University of Valencia. Yet, as the editors emphasizes, this is not a proceedings volume. All essays are written exclusively to appear in this volume. Most contributions will likely prove valuable for anyone researching matters in the intersection of agency, freedom, moral luck and responsibility. That said, it is less suitable as an introduction to the field since it lacks thorough introductions to many of the debates, problems, and positions under discussion.

The Metaphysics of Agency

The first part of the volume consists of four essays, the first two concerns intentional agency broadly, while the other two focuses on free agency more specifically. “The argument from slips” is the opening chapter. Here Santiago Amaya makes a comparison between perceptual illusions and action illusions. The former should be well known, but the latter are rarely discussed. According to Amaya slips are action illusions: “situations in which well informed, willful, and competent agents fall short of acting as intended”. The slip is a common mistake: you put the paper in your mouth and throw away the chocolate; you try to open your front door with the car key etc. Before you notice the mistake, it feels as if things go according to your plan. The kernel of Amaya’s argument is that just as the classic argument from illusion is an attack on ‘naïve realism’, the argument from slips can serve as an attack on what he dubs ‘naïve rationalism’. A position inspired by Ludwig Wittgenstein and G.E.M. Anscombe. This essay offers a refreshing new line of thought to philosophy of action. The argument from illusion has had an enormous influence on the philosophy of perception. The argument from slips could be a game changer, potentially provoking as much debate in philosophy of action as the argument from illusion has in the domain of perception.   

In “A Gradualist Metaphysics of Agency” Jesús H. Aguilar and Andrei Buckareff argue against the view that being an agent is an all-or-nothing matter. Instead, they assert, agency comes in degrees. All agents have in common that they are disposed to react in certain ways under certain circumstances. In the minimal sense, the authors argue, “a state exhibits some degree of intentionality to the extent that it is: (1) about, for, or directed at something; (2) directed at things that are present or may not be present; and (3) may exhibit indeterminacy with respect to what it refers” (p. 33). They denote this minimal sense of intentionality ‘quasi-intentionality’, and argue that chemicals may exhibit such intentionality since they are disposed to react when they come into contact with different properties of other objects. Aguilar and Buckareff make it clear they regard the requirements of quasi-intentionality as necessary but not sufficient for full agency. However, they leave the reader with the question of why we should accept a conception of intentionality that entails that chemicals have intentions. It seems problematic to abandon the idea that intentions involve at least some minimal mental activity on behalf of the agent. Instead, it would suffice to say that a state exhibits some degree of disposition if it fulfils all three requirements. Further, the authors argue that complexity of thought and richness of mental life comes in degrees, something which plausibly would entail that also agency comes in degrees — at least if you think that thought or mental processes are a constitutive element of agency. On this view, “rational agents are just the most complex agents that exists along a continuum with other types of agents” (p. 42).

The title of Michael McKenna and Chad Van Shoelandt’s essay “Crossing a Mesh Theory with a Reasons-Responsive Theory: Unholy Spawn of an Impending Apocalypse or Love Child of a New Dawn?” pinpoints the contents of the essay as well as it describes the authors own conviction. The authors argue that reasons-responsive theories can tell us something about the outer nature of free agency. On such accounts, an agent acts freely when the action originates in the deliberative mechanism of the agent; a mechanism that is sensitive to an appropriate pattern of reasons (see also Fischer and Ravizza, 1998). Mesh theories on the other hand can tell us something about the free agent’s inner nature. The idea is that an agent acts freely when there is a well-functioning harmony between the agent’s motivational states for that action, and some other relevant attitudes of her (see for instance the hierarchical and desire-based theory proposed by Frankfurt (1971). The authors are not prepared to commit to this hybrid view. Yet they find it persuasive since mesh theories can account for the problems arising from reasons-responsive theories, and vice versa. Even though the authors claim that you could insert your favourite mesh-theory as well as your favourite reasons-responsive theory, it seems that there is more work to be done in fitting these theories together. Could for instance Fischer and Ravizza’s requirement that the mechanism must be responsive to some moral reasons be combined with Frankfurt’s desire-based account? (Consider remarks made by Gary Watson, 1975).

In “Classical Compatibilism and Temporal Ontology”, Pablo Rychter argues that the following three views are compatible:

Eternalism: Objects both from the past and the future exist just as much as present objects.

Dual ability: Some agents are able to sometimes do other then they actually do.

Determinism: The laws of nature completely determine, together with past states of affairs, all future states of affairs. (p. 66)

He starts out by arguing that an eternalist can ascribe to ‘dual ability’ since the future part of the eternal universe could be different, given that other free choices had been made. To use Rychter’s example: “your having fish [for lunch] tomorrow is ‘already there’, waiting in line to become present. But […] it is already there because your free decision to have fish is also already there, just one place before in line” (p.69). A more pressing problem for compatibilists comes into picture if we add that the universe is also determinate. It seems then that the universe cannot contain another set of choices. Hence, it seems that the three views are incompatible. However, Rychter argues, this is only true if we also think that the laws of nature that determine the universe can be found in their entirety in the past. Instead, he appeals to a Humean conception of the laws of nature; the view that they are determined by “local matters of particular facts, just one little thing and then another” (p. 74). On this understanding, “we should see the laws as a result that follows from particular events, among which no necessary connection holds” (p. 74). The claim is not that we cannot know about the laws of nature other than from studying particular events, it is deeper than that. It is the claim that there are no underlying laws of nature that could explain why particular events turn out the way they do. There are no such necessary connections. We could question whether Hume would approve of this alleged Humean conception of the laws of nature. On a standard interpretation of Hume, his legendary skepticism was epistemological. We can never prove nor perceive the necessity of the laws of nature, only the regularities that they entail. Yet, Rychter needs more than epistemological scepticism to prove his point. Besides this exegetic consideration, there seems to be other questions that demand clarification. For instance, is the idea that agents can influence the laws of nature (particular to the situation) by exercising their free will? This seems necessary if agents are able — as a result of exercising their will — to sometimes do other then they actually do. Alternatively, is the idea that the laws of nature particular to a certain situation vary randomly? In that case ‘dual ability’ would be interpreted as saying that it is possible that agents sometimes do other then they actually do. Neither of the alternatives seems attractive. The first one is metaphysically demanding, and the second one seems to invite indeterminism, something which is problematic since we also assume that determinism is true. Finally, if the laws of nature are particular to the situation, why do they show irregularity in cases of agency, but regularity in cases of nonagency? Think of a case where a billiard ball strikes another. Hume, who stands at the side, notices that the other ball moves in another direction than he predicted — a way that must stem from a different set of laws of nature than he is used to. Rychter’s point of view on the laws of nature entails that such events are possible. Yet, such an unpredicted event would surely surprise Hume (as well as most of us). Arguably, such events are rare. However, when it comes to free agency, Rycher’s view entails that each time an agent could do otherwise, the laws of nature shows irregularity. How can this relation between free agency and the laws of nature be explained?

Responsibility and Luck

The second part of the book consists of four different essays. The first two essays defend libertarianism against the Luck Objection. More specifically, they defend the libertarian idea that agents sometimes exercise free will and that free will is logically incompatible with a deterministic universe. The luck objection states that if indeterminism is true, then undetermined decisions and actions are bound to be arbitrary or random rather then rational. Hence, which decision an agent makes is a matter of luck rather than free choice. In the first essay “Reasons and Freedom”, Carlos Moya argues that causally undetermined choices can be rational, especially focusing on the claim that deliberation is a matter of weighting reasons, not just weighing them. Further, he argues that rational choice does not require that the reasons on which choice is based are better than any alternative. It is often enough for rational choice to choose for good reason. The reasons need not be better than any alternative. In the second essay, “On the Luck Objection to Libertarianism”, David Widerker and Ira M. Schnall argue that proponents of the luck objection fail to sufficiently appreciate the very notion of leeway-libertarian free act. In asserting this, they especially focus on three different versions of the argument proposed by Alfred Mele, Neil Levy and Van Inwagen respectively. In particular, Widerker and Schnall argue that an action can be undetermined and still under the agent’s control. It is not necessarily a matter of luck that the causal gap left by indetermination becomes settled. Instead it could be that the agent steps in and exercise control over what she does. This argument could be seen as a more general variation of the argument Moya proposed in the previous essay.

In “Moral Luck and True Desert”, Sergi Rosell defends the claim that moral luck is an undeniable everyday phenomenon against what he calls ‘the Global Case against Moral Luck’. This case starts out from the control principle, that an “agent A is morally responsible for x only if, and to the extent that, she has control over x” (p. 117). Add to this the corollary “[p]eople ought not to be morally judged differently if the only differences between them are due to factors beyond their control” (p. 117). Considering different cases where the intervention of luck is progressively more distant from the action and its moral outcome, Rosell argues that the proponent of the Global Case against Moral Luck ultimately reduces the basis for desert either to the agent’s given constitution, or to a bare self with no properties at all. Further, Rosell argues that the former option involves abandoning the Control Principle since you cannot be in control over your given constitution. On the other hand, the latter option empties the notion of true desert since there is nothing left to base the true desert on (Michael J. Zimmerman, 2011, does not only bite this bullet, but uses it as one major argument against the possibility of justification of punishment). Rosell concludes the paper by claiming “the idea that what ultimately matters is only what exclusively depends on the agent, becomes a meaningless view, since it happens that, in the end, nothing exclusively depends on the agent” (p. 131). He doesn’t draw this argument to its ultimate conclusion, but taken seriously, his argument does not only defend the phenomenon of moral luck. It also — if true — refutes the control principle of moral responsibility.

Carolina Sartorio explore some aspects of moral luck in “A New Form of Moral Luck?”. She focuses on cases where our moral responsibility appears to depend on whether other responsible agents contribute to the unfolding of events. She discusses a rich variety of cases. In the first one, the goddess Diana induces a specific genetic mutation to the unborn child Ernie, giving him certain innate genetic dispositions. “Diana knows that, partly due to those dispositions and the deterministic laws, 30 years later Ernie will murder his uncle to inherit a fortune. Still, when Ernie murders his uncle 30 years later, he satisfies all the standard compatibilist conditions for freedom” (p. 135). In a similar case it is not Diana, but a flash of lightning, that causes the genetic mutation. Sartorio argues that the intuitions in these two cases are very different. In the first case, Ernie doesn’t seem blameworthy for his action, while in the second case he does. Yet, the only difference is that in the first case, an intentional agent plays the role that a natural force plays in the second. Another pair of cases where degree of responsibility seems to depend on whether other agents contributes to the unfolding of events is:

CASE 1: I want an explosion E to occur. I have good reason to believe that pressing button A will trigger an explosive that will result in E. I press A, and E occurs (p. 140).

and:

CASE 2: Three buttons (A, B and C) need to be pressed for E to occur. Two other agents independently press B and C while I press A (each of us knew about the buttons, had god reason to believe that the other buttons would be pressed, and acted with the intention that E occurs). E occurs (p. 140).

We can assume that the explosion has a morally bad outcome. Here, Sartorio argues, there are only two sensible options, either am I as responsible for E in case 2 as in case 1, or I am less culpable, for instance only a third as culpable as in the first case. However, this is not necessarily a case of moral luck as Sartorio thinks. You could agree that A has different degree of responsibility in case 2 than in case 1 without accepting that this has to do with moral luck. You could argue that A’s different degree of responsibility is due to the different intentions involved in the different cases. In case 1, A has an individual intention to cause E. Conversely, in case 2 A, B and C has a joint intention to cause E. Given the limited space of this review we cannot go into further detail, but it seems that the best explanation of the differing degrees of responsibility in some of the cases that Sartorio discusses is not that they are instances of a new form of moral luck.

Responsibility and Blame

According to the Principle of Alternate Possibilities a person is morally responsible for what she has done only if she could have done otherwise. The principle has been subject to much discussion since Harry G. Frankfurt claimed it was false (1969). Helen Steward’s “Helping it” is a contribution to this debate. Her focus is not merely Frankfurt’s account, but how we can best formulate the control intuition we have in cases of blameworthiness. More specifically, she claims that everyday locutions like “can’t help it” are a good way to formulate it. Her rationale for this is that these locutions offer a significant way in which the control intuition can be expressed. Consider a Frankfurt-type case. In this sort of case a crucial role is played by an agent with super powers who is able to intervene in other agents’ neurophysiology. Suppose a person, Jones, deliberates about whether to shoot the President. After some reflection he chooses to do it. Had he not made this decision of his own free will, the agent with super powers would have intervened and brought it about that he nonetheless killed the President. Frankfurt argues that Jones is clearly blameworthy in the case in which the intervener does nothing, but given his presence Jones could not have done other than shoot the president. Contra the Principle of Alternate Possibilities, we have a blameworthy agent who could not have acted differently. Steward’s question is whether Jones, additionally, couldn’t help shooting the President. Certainly, he couldn’t have avoided doing it, but according to Steward there is an alternative reading. True, Jones couldn’t help shooting the President as such, but it isn’t the case that Jones couldn’t help it: “It would be mighty strange, I think, to say of Jones, who in fact of course shoots the President calmly, coolly and collectedly of his own free will — ‘Oh, yes — Jones shot the President. But he couldn’t help it. There was this counterfactual intervener guy around, you know'” (p. 155). Steward argues for her case through more examples and details than we can go into here. She has provided a novel and exciting contribution to the ongoing debate about control in cases of blameworthiness.                   

In a country like Colombia that is overwhelmed by violence and war, there are people who cannot forgive despite the fact that they are deeply convinced that they ought to do so. In “Ought without Ability”, Carlos Patarroyo finds it wrong to conclude that these people just are mistaken, or that they are thinking of an ideal, prima facie and personally detached meaning of ‘ought’. He defends the possibility of there being a morally relevant meaning of ‘ought’ that does not imply ‘can’; a meaning that is not prima facie, ideal, non-binding or situational. In the spirit of R.M. Hare and Erasmus of Rotterdam, the discussion proceeds with understanding ‘ought’ as an imperative. Patarroyo makes use of Luther’s argument that a doctor could issue an order to a patient that is perfectly intelligible even though the doctor knows that the patient cannot comply. Hence, he concludes that commands that are agent-binding but cannot be obeyed make sense.

Derk Pereboom’s “Omissions and Different Senses of Responsibility” seems to have two purposes. First, to show that you can be as morally responsible for the outcome of decisions not to act and for failures to decide as you are for decisions to act. Second, to argue that there are senses of responsibility that do not involve basic desert. Yet, these two purposes are interrelated. Pereboom argues that the senses of responsibility that do not involve basic desert predominate in cases of decisions not to act and failures to decide to act. Pereboom start out criticizing Carolina Sartorio’s (2005; manuscript “The Puzzle(s) of Frankfurt Style Omission Cases”) claim that there is an asymmetry between responsibility for the outcomes of actions and responsibility for the outcomes of omissions. Sartorio considers a case where Frank sees a child swimming in a pool. Frank wants the child to die, so he jumps in and pushes the child’s head under water until he drowns. In a similar case, “Frank notices that the child is starting to drown. Since he wants the child to drown, he decides not to jump in” (p. 180). Sartorio argues that in the latter case, Frank is responsible for his decision but not for the outcome, hence the asymmetry. Pereboom considers cases where the agent (Frank’s counterpart) fails to make a decision. Plausibly, this won’t happen in cases where a child is drowning (because of the gravity of the situation), but it could happen in situations where harm is less severe. Such an example could be:

Stranded Motorist: “Frank is driving along a lightly-travelled highway thinking about Frankfurt cases. He notices a motorist struggling to replace a flat tire. After momentarily being distracted by the thought that the motorist would benefit from help, Frank resumes thinking about Frankfurt cases without ever deciding not to stop to help.”

(p. 182)

Since absences can be causes on Sartorio’s (2015, 2016) view, this would be a case where Frank by absence of decision caused the continued struggle of the motorist, and therefore he is responsible for this outcome. It would however be odd, Pereboom argues, if the responsibility-assessment were different in cases of failures to decide and cases of deciding not to act. He therefore concludes that we can be morally responsible for the outcome of decisions not to act. Yet, instead of drawing this conclusion, you could give up the premise that absences are causes. As Sartorio herself notices “absence causation is notoriously ‘explosive'” (Sartorio, 2016, p. 33). For instance, “the absence of an alien attack on Earth (one that would have wiped us all out in seconds) is one of the causes of Frank’s making his choice” (Sartorio, 2016, p. 34). To further illustrate this, we can consider another example Pereboom discusses: “Matt, a first-time parent living in an especially safe environment, whose first encounter with danger to his child is one in which she is seriously injured” (p. 189). It could be a case where Matt leaves his child in the car on a hot day, gets preoccupied, and returns to the car to find his child suffering severe hyperthermia. Pereboom argues that it is appropriate to blame Matt as well as Frank (in the stranded motorist example) since it is fitting to believe that their lack of vigilance was bad, and fitting to desire that they would not have failed in their respective ways. Yet, if you abandon the idea that absences are causes, you will see an important difference between the two cases (besides the gravity of harm). There is then no relevant sense in which Frank has caused the continued struggle of the stranded motorist, but there is a relevant sense in which Matt has caused his child’s suffering. After all, Matt put his child in the car and left her there. It seems perfectly conceivable to argue that we should put the suffering of the child — the harmful outcome — on Matt’s moral ledger, while we do not put the suffering of the stranded motorist on Frank’s moral ledger. In the latter case, we could still blame Frank for the failure to decide on the grounds mentioned earlier, but not for the outcome in itself.

The next two essays concern Michael McKenna’s recent book Conversation and Responsibility (2012). In his book, McKenna argues that blame is more easily justified than the moral responsibility skeptic might think. One important argument for this claim is that justification of blame doesn’t necessarily entail the justification of punishment, which is harder to justify. In “Moral Responsibility Skepticism: Meeting McKenna’s Challenge”, Neil Levy argues that blame in fact do harm us seriously. Therefore the difference in moral justification between blame and punishment is lesser then McKenna thinks. Another of McKenna’s (2012) arguments rests on the idea that blame is non-instrumentally good. A Contractualist-flavoured consideration lingers at the background; only someone who cares about moral relations can be exposed to harms distinctive of blame. For instance, if you do not care about interpersonal relationships, having them impaired won’t harm you. Against this, Levy argues that you can be harmed by being blamed even though you do not care about the noninstrumenal value of relationships since these relationships can be instrumentally valuable. Further, Levy argues that the value of being a member of a moral community cannot entail that blame itself has noninstrumental value. Rather, the convention of blaming comes at a moral cost. To the extent blaming is justified, it is so all things considered. In the next essay “In Defense of a Challenge to Moral Responsibility Skepticism: A Reply to Levy” McKenna distinguishes between the harm that blaming entails — which is not good — and the value of concern for others and morality  — which is good. He argues that when a blameworthy person is harmed because of blame (and not only since the relationships are instrumentally valuable) this involves an expression of her concerns for others and morality. Yet, it seems to us that this argument does not amount to saying that blaming is noninstrumentally good. What is instrumentally good is rather the agent’s concern for others and morality. Moreover, McKenna argues that skepticism about blaming must involve more than saying that blaming comes at a moral cost. For instance, most non-skeptics would agree that morally justified blame comes at a cost. It is enough for a non-skeptic to regard blame as justified all things considered.

Responsibility and Relationships

According to what Constantine Sandis in his “Motivated by the Gods: Compartmentalized Agency and Responsibility” dubs the Homeric Principle one is, “in acting against one’s intention or desire […] momentarily possessed by something alien to oneself and is therefore a victim rather than a culprit” (p. 211). Whereas it is common in Greek tragedy that deity is blameworthy for the fate of the protagonists, it is not common in contemporary everyday life to hold Gods responsible for one’s actions. It is, however, a commonplace that we sometimes fail to be ourselves, for example in times of sleep deprivation or stress. When this happens we act contrary to our intentions and desires. We do not approve of those actions and consider the parts of us that led up to them to misrepresent our true self such that it makes sense to claim less responsibility for them. Sandis questions this view, points to its weaknesses and suggests that even though we might bear less responsibility in these cases, it is simply wrong to think that they have absolutely nothing to do with us (p. 223). Sandis’ text should be of interest to a larger audience in addition to philosophers. Theologians are likely to take an interest in the discussion whether deity can be blamed for human actions. And the use in the text of vivid examples from film and literature is likely to intrigue people from the field of comparative literature as well.