04 Jun Is Any Libertarian Account Of Free Action Adequate?
In this essay, I answer the question which, if any, libertarian account of free action is adequate. I take free action to rest on two pillars, namely the agent’s ability to do otherwise (required for moral responsibility) and the agent’s being the ultimate source of his actions. Libertarian accounts of free action hold that at least some actions do not have a sufficient cause, where “cause” is understood in the non-agential sense. In order to see if any libertarian account is adequate, I examine several approaches from contemporary literature, categorized into event-causal, non-causal and agent-causal theories. My conclusion is that Daniel von Wachter’s view is the only adequate approach on the marketplace, although it leaves some questions unaddressed. I will identify those questions and fill in the gaps with my own theory.
I Clarifying the relevant terms and the question
Libertarian accounts of free action are incompatibilist in nature. That means, they take determinism and free action to be incompatible; in contrast to hard determinists who deny free action, libertarians take determinism to be false. Determinism is the doctrine that every event is conditionally necessitated by prior events and circumstances, or, in other words, that every event or action has a sufficient cause. Because of this, many philosophers have thought and still think that indeterminism is required for a libertarian account of free action (Clarke 2011, 331). We shall see, however, that indeterminism, in most accounts, destroys freedom rather than fosters it; also, it can be shown that indeterminism is not even required for free action (Von Wachter 2009, ch. 7). Libertarian accounts can be categorized into three groups. Event-causal accounts take it that the actions of an agent are caused by prior events and circumstances, but not deterministically; hence, they require indeterminism at some point. Non-causal theories claim that a free action is uncaused. Agent-causal approaches take the agent to be the cause of his actions, explicitly precluding that prior events are the (full/sufficient) cause of an action.
Free will is generally understood as the kind of control over actions an agent needs for moral accountability. This includes the ability to do otherwise as well as the ability to originate one’s actions, i.e. to be the ultimate source of them.
A free action is one that issues from a free will and which is uncoerced.
The present question seeks to find out whether any libertarian account renders free action adequately. That is, in order to be adequate, a libertarian theory of free action needs to respect the requirements of free will outlined above. If, for example, a libertarian account turned out to deny agents to be the ultimate source of their actions, it would not count as adequate. As we shall see, a great problem is the chanciness that many libertarian accounts adduce by relying on indeterminism, thereby robbing the agent of any control over his actions. Other accounts, while seemingly solving the chance problem, leave no or too little room for the agent and thus for someone to exercise the required control.
II Event-causal accounts
As an exemplary of a sophisticated event-causal libertarian theory, I examine Laura Ekstrom’s “preference theory” (e.g. Ekstrom 2011). The centerpiece of her account is the notion of “preference”. By preferences she understands desires that have passed a process of critical evaluation with one’s conception of the good (ibid, 371). The formation of such a preference, she maintains, is an action. Here is where indeterminism comes into play in Ekstrom’s theory: preferences are brought about indeterministically. Now a preference thus acquired deterministically leads to a decision or performance of an action, which she then considers to be free:
The resulting decision output, the preference, when indeterministically caused and noncoercively formed, is authored by the agent, since it is formed by her for reasons that justify and explain it, and its claim to being authentic is not defeated by the objection that she formed it because she had to, because it was causally necessitated by the past and the natural laws. (ibid., 373)
The problem with this approach is that an action is declared as free even if it is brought about by events none of which is a free action. Now, given the deterministic connection between preference and action and the indeterministic formation of the preference, it seems that the action is not up to the agent. An indeterministic causation consists in the effect not being causally necessitated by the cause; the preference might or might not have arisen. The agent seems to have no control over the outcome during the preference formation or during the performance of the action (which is determined by the preference). Ekstrom (ibid., 375-76) attempts to salvage her account by saying that the chance element in the indeterministic causation of a preference should be understood in a probabilistic sense, i.e. that there was a probability 0 < p < 1 of the event occurring and a probability q = 1 – p of the event not occurring. Thus, she maintains, one need not understand chance as a weird force (“Chance”) determining the outcome or as the mere absence of purposiveness. Ekstrom holds that her account yields exactly what is required for purposive agential control, namely considerations (indeterministically) entering the agent’s mind and subsequent deliberation. As I understand Ekstrom, the chance would then consist in it not being determined which considerations enter the agent’s mind. The subsequent deliberation cannot be chancy, as Ekstrom insinuates with the use of the term “purposive”.
However, this explication does not fix the account. Granted, considerations popping up indeterministically do not destroy free will, as long as the agent has the ability to pick either as her preference. But here becomes obvious what is lacking in Ekstrom’s account. According to her, the agent is her preferences along with her capacity to form them by reflective evaluation. But a “package of preferences” deterministically leading to actions together with reflective evaluation is not enough for free will. It takes the agent to set his will on one of the options that reflective evaluation brings to light. But no such agent is in the offing. If we assume “agent = preferences”, we end up by concluding
- The agent = (a bundle of) preferences + reflective evaluation.
- Free action requires that the agent picks one preference upon reflective evaluation.
- Therefore, a free action requires that a bundle of preferences picks one preference.
This seems absurd. Therefore, Ekstrom’s account is inadequate for free will.
What about other event-causal accounts? They all depend on indeterminism in a causal chain of events. Randolph Clarke and Justin Capes (Clarke and Capes 2017) nicely sum up the problems all event-causal accounts face due to this assumption:
If this strategy [of relying on indeterminism to salvage alternate possibilities] succeeds in showing that the required indeterminism would not undermine responsibility, it leaves unaddressed the charge that the requirement is superfluous, that it secures nothing of value that could not exist in a deterministic world. And it is hard to see how this charge can be answered. (21)
III Non-causal accounts
Non-causal accounts as Carl Ginet’s (e.g. Ginet 2007) hold that a free action is neither caused deterministically nor indeterministically, but uncaused. As this seems puzzling, Ginet makes his claim clearer by assuming that the “causally simplest events that count as actions are mental events, like decisions and volitions” (ibid., 244). These mental events, Ginet claims, are uncaused in free actions. He grants that the subsequent (perhaps bodily) action is in turn caused by those mental events, but maintains that the action as a whole is still uncaused, because its first stage triggering the subsequent stages was uncaused (ibid., 245),. In order to distinguish his view from causal pictures, he states that “given that an action was uncaused, all its agent had to do to make it the case that she performed that action was to perform it” (ibid., 247). Whereas on a causal picture: “For any event e, S made it the case that e occurred only if S caused e to occur.” (ibid., 246; emphasis added).
Two main objections have been raised against this. The first one is that if an action is uncaused, it must be a matter of chance or luck, thereby stripping the agent of all control and accountability. But Ginet correctly answers this by maintaining that the luck objection applies to indeterministic/probabilistic causation only; any probabilistic understanding of chance does not apply to uncaused events. Only if an action like telling the truth is indeterministically caused, it makes sense to speak of a probability applying to it. Let’s assume that to action A a probability of p = 0,57 applies. If A is indeterministically caused, this would mean that on 100 instances in which prior events and circumstances conducive to A obtain, only in 57 of those 100 instances A occurs, this ratio being a law of nature. In other words, we could predict that of 100 instances 57 lead to the occurrence of A. If A is uncaused, however, we could not predict at all the outcomes of the 100 instances. Hence, it makes no sense to speak of a probability, if by that a law of nature is meant; we could just say retrospectively that of 100 instances, 57 led to the occurrence of A, not being justified in contending that the next 100 instances will exhibit the same ratio.
The second objection takes up motives and reasons as necessary explanations for actions. It goes as follows (ibid., 251):
- an action could have been up to the agent only if it has an explanation in terms of the agent’s motives or reasons; and
- its being uncaused would preclude such an explanation.
Ginet replies by arguing that explanations need not be causal explanations. Therefore, even if a reason is the explanation for an action, the reason need not have caused the action. The relation between reasons/motives and actions is then not a causal one, but rather an intentional one: “[M]y reason for doing A was that I wanted to obtain B and believed that by doing A I would obtain B.” (ibid., 251) A similar view is proposed by Thomas Pink (2011).
I take it that the non-causal approach is robust. Its weakness, as I see it, lies rather in what it does not say than in what it says. Ginet speaks of the agent making it the case that he triggers an action by non-causally bringing about a volition. Now to me, the question arises who or what the agent is. Ginet himself is not clear about whether he views the agent as a substance or not (Ginet 2007, 245). He maintains that the question of the agent’s ontological status is irrelevant to the question whether an action can be uncaused and up to the agent. I think, contra Ginet, that it is relevant. As we have seen, the non-causal view is still committed to actions having an explanation. That means, proponents will reject actions as being due to ontic chance, i.e. having neither a cause nor an explanation. Now the agent figures prominently in Ginet’s and Pink’s theories; it is his reason-based intentionality that makes actions happen. To me, this cries out for an account of what sort this being – that can bring about volitions “ex nihilo” – is. As we shall see, an account classically attributed to the agent-causality camp might do, even if Ginet will surely reject the term “causation”. But in my view, how we label the relation between an agent and his volitions (or tryings, choice-events, undertakings and whatever else they have been called) is a matter of terminology and not one of metaphysics. Let us now see what agent causalists have proposed concerning free actions.
IV Agent-causal accounts
Agent causality means that it is true that the agent brought about the action, and “it must not be the case that either what the agent causes or the agent’s causing that event is causally determined by prior events” (Clarke and Capes 2017, 23). What exactly it means that the agent brought about the action is a matter of ongoing debate. I will look only at accounts that establish metaphysical theses. Roderick Chisholm’s (1976) approach, for example, seems to analyze agent causality merely semantically, holding that a statement of the following form:
(ACS) X did Y, where X is a free agent
cannot be reduced to statements containing only event causation without loss of meaning. In fact, Chisholm designates even clearly unfree actions as caused by the agent, just because propositions about those actions cannot be reduced to event causality statements without loss of meaning. Hence, I take this approach, as Daniel von Wachter points out, to be a “child of the linguistic turn” (Von Wachter 2003, 8), unsuitable to account for free actions in a libertarian sense.
Let us turn to Randolph Clarke’s account as a first truly metaphysical thesis about agent causation. Clarke defines agent causality as follows (Clarke 2003 as quoted in O’Connor 2011):
- whatever action is performed will be caused by the agent,
- a reason n will cause an action A1 only if the agent causes it, and
- the agent will cause an action only if a corresponding reason causes it.
Or, graphically represented:
This means that the agent is a co-cause alongside his reasons. The reasons themselves have a tendency of their own to produce the action. But that is not a good way of construing agent causality, because it ascribes too much causal power to the reasons: it seems that even without the agent, the reasons have a certain chance of bringing about the action. O’Connor (2011) mends this by saying that a reason increases the agent’s propensity to do action A. But, importantly, „the agent is the sole causal factor directly producing her intention to A (not a co-cause along with her reasons, as in Clarke’s view), but her deliberation and activity take place within an internal context (including her total motivational state) that has probabilistically delineated causal structure.“ (ibid., 317). This seems to aptly capture our intuitions that we are the ultimate sources of our actions, but that we do act for reasons, even if those reasons do not determine our actions. Still, two questions remain open:
- Our actions depend on brain events which are subject to laws of nature. We don’t want to say that we as agents are also subject to laws of nature (thereby flattening our freedom) and we don’t want to say either that we as agents can break the laws of nature when we act. So how can agent causality be possible without breaking the laws of nature?
- What kind of being is an agent?
The first question might be addressed by introducing indeterministic processes into agent causality. But as I have shown above (section I), this is a Trojan horse for the agent causalist. If it is just a matter of chance or probability whether neurons fire or not, the agent has no control over his actions. Instead, we might try a picture analogous to Alvin Plantinga’s (2011) Divine Collapse-Causation (DCC). DCC claims that all collapses of quantum wave-functions are caused by God, which Plantinga takes to be God’s “open crack” to interact with the world beyond creation and conservation. Now we might transfer this idea to the human brain and say that it is the agent who causes at least some collapses of indeterministic wave-functions in his brain, thereby triggering a causal chain that leads to a certain action. This looks like a feasible theory, but it requires an agent as an irreducible and perhaps immaterial substance, which faces certain challenges (see section V).
Robert Kane goes a different way. He takes indeterministic processes to be a hindrance rather than a positive causal contributor to deliberation and action (Kane 2011, 393). In Kane’s picture, an agent faced with a decision runs parallel efforts to do both actions; indeterminacy in the brain processes is a hindrance to one option or both; the option effectively instantiated is the one whose hindrances the agent overcame. What is not entirely clear from Kane’s account is what ultimately settles between two or more options. He might be read as claiming that it is indeterminacy that hinders one option so strongly that the other one is materialized. But to me that sounds like a great loss of control over the outcome and in moral responsibility. Take the example Kane himself gives: a businesswoman walks to a meeting crucial for her career and suddenly sees the injured victim of an ambush lying in the bushes. Now according to Kane, two efforts run parallel in the businesswoman’s mind, namely (i) trying to ignore the victim and get on to her meeting and (ii) trying to stop and help the victim on pain of missing her meeting. Which one she will do ultimately depends on which of the brain processes corresponding to (i) and (ii), respectively, will be less hindered by indeterminacy. Let’s assume the businesswoman has recently become aware of her selfishness and desperately wants to improve her character. The victim offers her a chance to do so. On Kane’s account, if she fails to help the victim, she might excuse herself by attributing the outcome to indeterministic brain processes; after all, she might say to herself, I tried to help the victim, but of course had reasons to ignore him, so after all the career-oriented action won over. Perhaps Kane would reply that if prior to the incident she had made a self-forming action (SFA) towards being more altruistic, the outcome would be different; one would then have to praise her rather for the SFA than for the actual help. But to me, this is an unrealistic picture. Very often, people report being tempted and having given in to temptation despite prior will-setting towards moral praiseworthy actions.
Alternatively, Kane might mean that in the settling no indeterminacy plays any role, but the decision of the agent (presumably guided by reasons). If so, the question arises who is performing the action against the indeterministic “harassing fire”. Kane calls the agent a persistent substance, but gives a reductive account of it. He claims that the agent can be reduced to “states of affairs, events and processes involving it” (ibid., 396) and that they are “systems (…) in which new emergent capacities arise as a result of greater complexity or as the result of movement away from thermodynamic equilibrium toward the edge of chaos.“ (ibid., 396). What “emergent capacities” really arise in organisms due to a movement away from thermodynamic equilibrium remains obscure, not to speak of the possibility of mental capacities (such as will-setting) from pure matter. At any rate, the whole idea perspicuously sounds like epiphenomenalism, against which there are strong and, in my view, decisive arguments (see e.g. Swinburne 2013). I shall offer some arguments against an emergentist view of the agent in section V and conclude that Kane’s account, although bringing up some legitimate points (parallel processing, SFAs), is not an adequate account of free action.
The most promising account on the marketplace is Daniel von Wachter’s theory of agent causation (Von Wachter 2003), Von Wachter 2009). He takes agents to be capable of bringing about what has been labeled in various ways “efforts” (Kane 2011), “volitions” (Ginet 2007), “endeavorings” (Chisholm 1976) or “tryings” (Swinburne 2013); his own term, to which I will also stick to, is choice events. Von Wachter uses this term to distinguish those events (also called by him initial events, because the initiate an action) from events that are part of physical processes. Importantly, according to von Wachter, choice events are not caused by prior events, but by the agent. He leaves open whether choice events are mental or physical events (presumably brain) events. Also, his view allows, but does not depend on, indeterminism. That is because he rejects classical determinism altogether and instead proposes a “directedness” theory of causality ((Von Wachter 2009, ch. 5; Von Wachter 2012) in which things together with states of affairs form a basis for a tendency towards another state of affairs that is “deterministically” realized unless another tendency incompatible with the first one prevents it. Thus, the choice event together with the brain matter form a basis B1 for a tendency T1 to action A (say raising one’s arm), which happens unless another tendency T2 incompatible with T1 (e.g. a cramp in the arm muscle) prevents it from occurring. No indeterminism is required to leave open alternate possibilities, but of course indeterminism might be adduced along the lines of Plantinga’s DCC (see above).
On that picture, Von Wachter rightly claims to have solved the dilemma of free action:
With choice events the dilemma of free action is solved. Free actions are neither caused deterministically nor are they uncaused or indeterministically caused. Choice events are the third way that avoids both horns of the dilemma. (Von Wachter 2003, 4)
Of course, his theory faces all the objections that have been raised against agent causation. I shall reply to some of them further below. First, I want to make clear my motivation for taking Von Wachter’s ideas further than he himself does. It should have become clear by now that one main problem with many libertarian views of free action is that they don’t posit the agent as an irreducible substance. Von Wachter takes it to be likely that agents are immaterial substances (Von Wachter 2009, ch. 7.7), but also views materialistic or reductionist conceptions of the agent as compatible with his theory. I think, however, that a reductionist conception of the agent faces insuperable problems, wherefore I submit that we should take the agent to be an irreducible, immaterial substance.
V The agent as an irreducible, immaterial substance
On my account, the agent is an irreducible, immaterial substance. The choice events this substance brings about are therefore mental events. Those events typically cause physical events in our body that lead to a certain action, but they might also cause other mental events without causing a physical event. Why do I think this substance dualist move necessary? It is because the alternatives, (i) emergentist views and (ii) materialistic/eliminativist views are extremely implausible. Let me qualify that.
An emergentist view like Robert Kane’s (2011) roughly accepts that the agent exists as a persisting substance, but that this substance is just an emergent phenomenon of more basic physical facts about the human body. Now there is widespread agreement that something like emergence exists, although our grasp of the notion is surprisingly vague (cf. Mumford and Anjum forthcoming). At any rate, to my knowledge there is no good argument why we should accept that mental properties can emergence from non-mental properties. We know of emergence of new physical properties from the co-working of physical parts (e.g. water has completely different properties than its constituent chemical elements hydrogen and oxygen). But there is no case in which clearly the composition of physical compounds led to an emergence of mental properties. To say that we humans are the best proof, because in us mental properties emerge from physical facts is of course question-begging. If one day scientists manage to build a human being from inanimate matter, and this human being upon completion exhibits mental properties, I will begin to give mental emergentism a chance – no earlier.
What about a materialistic account of the agent? Daniel von Wachter suggests that a materialist might say that it is the brain which brings about choice events (Von Wachter 2003, 4). But the materialist’s view will then be inconsistent. How can a brain bring about a choice event? Surely, the materialist will claim that choice events are just brain events. Granted. But how can a brain bring about brain events? Presumably, the materialist will be committed to laws of nature. If he takes them to be deterministic, the question arises which cause caused the choice event. The brain as a whole? This might mean “a plurality of other neurons”. But then it is open what makes those neurons fire, and we’re on to a vicious regress. The materialist might alternatively say that the choice event was uncaused, but then it is truly uncaused, not in Ginet’s sense above. Well, even if such “ontic chance” is possible, it is certainly not a good foundation for free will, because it seems to preclude any control over the action. Either way, it seems that the materialist cannot give a satisfactory answer as to how this agent can be free in the sense of bringing about choice events. As neither the emergentist can, I take it that the next most plausible candidate is substance dualism. Some challenges dualist accounts of agent causality face will be addressed in turn together with objections to agent causality.
The intelligibility objection contends that it is “unintelligible” or “mysterious” how an agent causes a choice event. Proponents conclude that a metaphysics containing agent causation is “panicky” or “extravagant”. But surely it is bad philosophy to label something phenomenologically so widespread as agent causation of choice events “unintelligible” and forget it. Just as with John Mackie’s “queerness” objection (Mackie 1991) against objective morality, it can be shown that the intelligibility objection rests on a double standard: If agent causation is unintelligible, so is event causation; but if the latter is acceptable (as most philosophers will agree upon), so should be the former. Is the objection effective at least against an immaterial agent? Isn’t it mysterious how an immaterial agent brings about changes in the physical world? Here, too, we have a parallel case that stands and falls together with immaterial agent causality: divine interaction. If such a thing is possible, then immaterial agent causality is possible, too. Of course, a materialist will presumably reject divine interaction together with immaterial agent causality, and it is perhaps not as obvious that divine interaction exists as that physical causation exists. At any rate, there is no conceptual impossibility for either divine interaction or immaterial agent causality, especially as determinism seems to be false (Von Wachter 2012). Furthermore, there is a plethora of reports claiming to have witnessed divine interaction. I cannot treat that topic more deeply, but let it suffice here that it seems at least reasonable to assume immaterial agent causality, even without knowledge of mechanism.
The datedness objection complains that a cause must exist before its effect, but there is no clear sense in which an agent pre-dates its choice events. I think this objection is mistaken from the outset, for several reasons. First, it takes a condition reasonable for event causation – the condition that a cause must exist before its effect – and transfers it unreflectingly to agent causation. But of course an agent is ex hypothesi not an event. Second, even if we grant something like Randolph Clarke’s claim that “a substance is in time only in that events involving it… are directly in time.” (Clarke 2003, 201-2, as quoted in O’Connor 2011), no refutation of irreducible agent causality takes place. If Clarke means that events are ontologically more fundamental than substances, it would be far from being accepted off-handedly and would need argumentative support. Maybe he rather means that a substance S exists at time t iff there is an event E at t involving S. But then E also exists at t only because S exists at t, and E is not more fundamental than S; the objection gains no ground.
The objection deflected for now, can we give a positive account of how an agent-substance can bring about a choice event? A good start is to say that some properties of the agent-substance ground its causal power to bring about choice events. But how this account fares totally depends on how “properties” is construed here. If it means that a property p of the agent S together with states of affairs B (e.g. the occurrence of a reason in S’s mind) necessarily causes an action A, there is neither freedom nor anything to do for S; in fact it seems that the postulation of an agent is superfluous. But that cannot be meant. I suggest that p be construed as the ability to bring about a choice event and p* as the property of influencing brain matter via a choice event. One might ask now what determines which choice events an agent brings about and when he does it, but that leads to the vast field of reasons and motivations for action which I cannot treat here. Let it suffice to say that reasons influence but do not determine an agent’s action and.
A last objection expresses the worry that positing an irreducible substance-agent unnecessarily tears down the uniformity of causal powers. Why, it is asked, should we posit an additional agent-causal power when we already have event causation that is applicable in every instance of causation? Surely this would be a metaphysically more parsimonious picture. I reply that event causation cannot satisfactorily explain every instant of causation, as we have seen. Therefore, we ought to assume an additional form of causality, even if this “bloats” our metaphysics. After all, Ockham’s razor only demands that we do not multiply entities and properties without need. In my view, the reason why such widespread aversion against irreducible agent-substances persists is that many philosophers are somehow committed to mechanistic worldview, even if they reject determinism. This is supported by the fact that many incompatibilist philosophers seem to regard indeterminism and non-causality as the only solutions to maintain free will. I take it to have shown that the idea of agents as immaterial, irreducible substances capable of bringing about choice events is coherent concept with great explanatory power.
Chisholm, Roderick. 1976. “The Agent as Cause.” In Action Theory, edited by M. Brand and D. Walton, 199–211. Dordrecht: Reidel.
Clarke, Randolph. 2003. Libertarian Accounts of Free Will. Revised. Oxford: Oxford University Press.
———. 2011. “Alternatives for Libertarians.” In The Oxford Handbook of Free Will, edited by Robert Kane, 2nd ed. Oxford University Press.
Clarke, Randolph, and Justin Capes. 2017. “Incompatibilist (Nondeterministic) Theories of Free Will (PDF Version).” Stanford Encyclopedia of Philosophy. https://plato.stanford.edu/entries/incompatibilism-theories/.
Ekstrom, Laura. 2011. “Free Will Is Not A Mystery.” In The Oxford Handbook of Free Will, 2nd ed. Oxford University Press.
Ginet, Carl. 2007. “An Action Can Be Both Uncaused and Up to the Agent.” In Intentionality, Deliberation, and Autonomy, edited by Lumer, 243–255. Ashgate.
Kane, Robert. 2011. “Rethinking Free Will: New Perspectives On An Ancient Problem.” In The Oxford Handbook Of Free Will, edited by Robert Kane, 2nd ed. Oxford University Press.
Mackie, J. L. 1991. Ethics: Inventing Right and Wrong. 49307th edition. London: Penguin Books.
Mumford, Stephen, and Rani Lill Anjum. forthcoming. “Emergence and Demergence.” In N.N. Taylor & Francis.
O’Connor, Timothy. 2011. “Agent-Causal Theories Of Freedom.” In The Oxford Handbook Of Free Will, edited by Robert Kane, 2nd ed. Oxford University Press.
Pink, Thomas. 2011. “Freedom And Action Without Causation: Noncausal Theories Of Freedom And Purposive Agency.” In The Oxford Handbook of Free Will, edited by Robert Kane, 2nd ed. Oxford University Press.
Plantinga, Alvin. 2011. Where the Conflict Really Lies: Science, Religion, and Naturalism. 1st ed. Oxford University Press.
Swinburne, Richard. 2013. Mind, Brain, and Free Will. Oxford: Oxford University Press.
Von Wachter, Daniel. 2003. “Free Agents as Cause.” In On Human Persons, edited by K. Petrus, 183–94. Frankfurt/Lancaster: Ontos Verlag.
———. 2009. Die Kausale Struktur Der Welt. Alber.
———. 2012. “Kein Gehirnereignis Kann Ein Späteres Festlegen.” Zeitschrift Für Philosophische Forschung 66 (3): 393–408.