If every action is caused solely by another event and a cause necessitates its effect, then there is no action to which there is an alternative (C).

The Libertarian, who believes that alternatives are a requirement of free will, responds by denying the right conjunct of C’s antecedent, maintaining that some actions are caused, either mediately or immediately, by events whose effects could be different, even if they were to recur under identical circumstances. We have here a denial of Laplacian Determinism (LD), according to which the condition of the world at any instant makes only one state possible at any other instant.¹ One prominent defender of this view, Robert Kane, holds that unless an agent’s neural mechanisms operated indeterministicly in forming her character she is not responsible for its manifestations.² This requirement is entailed by the principle of "ultimate responsibility" (UR) according to which an act is freely willed only if (a) its agent is personally responsible for its performance in the sense of having caused it to occur by voluntarily doing something that was avoidable and (b) there is no sufficient condition for its performance for which its agent is not personally responsible.³ The orthodox Compatibilist, adducing counterexamples to UR, argues that even if C and its antecedent are true, free will is possible.⁴ Ned Markosian has provided the Compatibilist with yet another response to UR-based objections to her view, wedding it to the notion of agent causation: here the deterministic causing of an act by its agent is necessary and sufficient for its being "morally free."⁵

Let us begin our assessment of Markosian’s compatibilist theory of agent causation or COMTAC by comparing it to the orthodox theory of agent causation (OTAC). On the this view, which also posits alternatives as a requirement of free will, C’s antecedent’s left conjunct is false because of cases in which an agent herself is the sole cause of her action. Moreover, in such cases the agent indeterministicly causes her action, being capable of doing something else under the circumstances.⁶ COMTAC, so as to so as to be consistent with LD, does not affirm the "ability to do otherwise" requirement. According to it, an agent caused act is not only determined but overdetermined, being the necessary consequence of both an exercise of its agent’s "active power" and mental events, specifically, the desires and beliefs explaining its performance.⁷ Thus, it would be incumbent upon COMTAC’s defenders to explicate the relationship between these causes. Markosian himself holds that the mental events that make up the motive of an agent caused act are both its mediate and immediate causes, immediately causing both the agent causing of the act and the act itself.⁸

Treating the agent causing of an act as an effect, as others have argued, is problematic; but I shall not press this point here.⁹ Instead, let us consider an apparent inconsistency in COMTAC: as its immediate cause, mental events must precede an agent causing, but, in overdetermining its effect, they and the agent causing must occur simultaneously. (One might ask instead, could an event be both an immediate and a mediate cause of the same action?) Thus, it seems that Markosian must either abdicate the idea that there are two immediate causes of an agent caused act or abandon the notion that the motive of an agent caused act causes its agent causing.¹⁰

Markosian could maintain his commitment to the counterfactual theory of causation while exercising the former option, for it would not be true in his example that (Imran’s) passing of the salt was not caused by (Yasmine’s) request(ing)" of it, were the latter not its immediate cause, as it would still be its mediate cause.¹¹ Regarding the latter alternative, Markosian believes that an agent causing is a change in its agent and, thus, an event, which requires a cause. But, an object changes only if it changes within a "quality space," that is, under a determinable and it is not clear that there is a quality space covering the case of an agent causing: what determinants would be involved here?¹² Going from not causing to causing an action seems to be a "Cambridge" change in an agent: her action is something of which she becomes the agent; there is no way in which she alters. (Compare: Imran’s brother’s son is someone of whom Imran became an uncle.)

There is a third option here, of course. COMTAC would remain consonant with the Compatibilist’s belief that every action satisfies LD were Markosian to abandon both claims, leaving an agent caused act causally unrelated to prior events yet still the necessary effect of an agent causing. Its motive could, thus, be taken as a mental event that, as Leibniz famously put it, "inclines but does not necessitate" its agent to cause a certain act to be performed, the determining factor being the agent’s exercise of her active power.¹³ Presently, I shall adduce further reasons, which are entailed by the Incompatibilist concerns that COMTAC is intended to address, for making this move. The question is: would an agent caused act with a mediate or an immediate causal connection to prior events satisfy UR?

As Markosian notes, one reason that the Incompatibilist gives for thinking that such an act does not meet this standard is that it was determined by events that took place before its agent was even born.¹⁴ COMTAC affords the Compatibilist the response that its agent is also a determining factor and is, thus, at least partly responsible for its performance, even if her agent causing of it was itself determined by primordial events. But, the Incompatibilist would grant this point to the orthodox Compatibilist, who attempts to "characterize the right (responsibility entailing) way for (a) causal sequence to flow through an agent."¹⁵ The Incompatibilist does not deny that standard Compatibilist definitions of ‘free will’ capture important senses of responsibility.¹⁶ Her complaint, though, is that such definitions fail to take into account certain "powerful intuitions" regarding responsibility- that is, their satisfaction would not entail the existence of free will in the ‘deep’ sense of UR.¹⁷ For example, on the view that I favor, responsibility is a matter of an agent forming and executing an intention to perform a certain act by exercising her will under conditions favorable to its proper functioning, which raises the issue of how an agent can be responsible for the exercise of a faculty with which she has been endowed.¹⁸ It is not clear that COMTAC, which does not require of a freely willed act that its agent be its ultimate source, satisfies condition (b) of UR either.

Anticipating this objection, Markosian constructs a counterexample to the claim that if something besides an agent’s exercise of his active powers make it physically necessary that he agent cause an act, then she is not responsible for that act.¹⁹ Here a student named Zane, who was already very likely to ‘always do the right thing’, becomes fully incapable of acting wrongly due to the influence of a "gifted and inspiring teacher." Should Zane go on to agent cause himself to rescue hundreds of children from a burning building, Markosian holds that he would deserve praise, despite his teacher’s influence having contributed to it being physically necessary that he would agent cause himself to act heroically. I think that the Incompatibilist would grant this point (saying that it is not clear that the Zane would not have acted as he did sans his teacher’s influence)- but only if it could be shown that Zane himself was ultimately responsible for having developed his character to the point where it only took a little more urging by a dedicated teacher to make him morally perfect. Markosian assumes that Zane satisfies this requirement, since his character arose from him having agent caused a series of physically unnecessary actions. But character development from undetermined exercises of one’s active power would be possible only if LD were false, which is not supposed to be the case, according to Markosian’s presentation of COMTAC, whereby an agent caused act is also the necessary effect of prior events. In other words, this example seems at odds with the stated purpose of COMTAC: to account for free will given the truth of LD. Moreover, COMTAC would be superfluous if such character development were possible. Unnecessitated character development is the basis of Kane’s Libertarianism. Again, Markosian can make his Zane example work only by making the case for an alternative to COMTAC: as Kane points out, why would a free will theorist bother defending a scientifically problematic form of causation if indeterministic event causation can be shown to account for free will?²⁰ Alternatively, Markosian’s handling of this case suggests that it would be best to modify COMTAC as suggested above: why not just have it that agent causing is in general unnecessitated by prior events if you are going to allow the agent causings that produce one’s character to be undetermined? In any event, we have been given no reason to think that COMTAC meets requirement b of UR.

Ditto Peter van Inwagen’s "Beta" objection, which is a formalization of this charge. If I have no control over either a law linking as cause to (distant) effects a primordial fact to my actions or that primordial fact itself, then I seem to lack control over my actions.²¹ Granted, I played a causal role in each one being performed, if COMTAC is true. But, on this view, they would have been performed sans my agent causing of them and, more importantly, the agent causing in each case had to occur given antecedent conditions and the laws of nature. From the Incompatibilist’s perspective, I have here no more responsibility entailing control over my conduct than I would be were my actions the necessary effects of unimpaired exercises of a faculty of mine over whose development I had no control. OTAC treats agent causation as indeterministic and causally independent of prior events precisely so as to forestall this objection.

LD is also said to pose an "indirect" threat to free will. If an act is determined, then its agent cannot do otherwise and if an agent cannot do otherwise, then her will is not free. Thus, if an act is determined, then its agent’s will is not free. Here LD does not directly eliminate free will, as with the above objection to Compatibilism; rather, it rules it out by rendering our actions unavoidable, so that none of them would satisfy condition (a) of UR.²²

Markosian deems "Frankfurt cases" abstruse and, thus, unlikely to convince Incompatibilists that free will does not entail alternatives. COMTAC, on the other hand, is supposed to provide the Compatibilist with a stronger response to this objection: for if an act is caused by its agent, then it is something for which she should be held responsible- even if it was unavoidable. But the Incompatibilist will have the same problem with an agent causing being unavoidable, which it would be according to COMTAC, as she would with the lack of alternatives to an act caused by an event. If an agent’s causing of a certain action were unavoidable and responsibility entails alternatives, then she would not be responsible for her agent causing of that act and, thus, the act itself. Again, the Incompatibilist will readily concede that there are senses in which an agent can freely will an act without it being avoidable. She will insist, however, that in the ‘deepest’ or philosophical sense an agent must have alternatives to what she is doing (or at least have had them to something she did to make what she is doing unavoidable) in order to be acting of her own free will.

In sum, COMTAC does not appear to be an improvement upon orthodox Compatibilism. Having converted to the former, a theorist would still be faced with UR-based objections to her position, in addition to now having to explicate and defend a scientifically problematic type of causation. In other words, a Compatibilist attracted by an agent-causal account of free will must be prepared to countenance actions lacking events as deterministic causes, if not abandon LD altogether.

1. John Earman discusses the status of LD within the Special and General Theories of Relativity in A Primer on Determinism (Dordecht: Reidel, 1986) pp. 170-198. Roger Penrose takes up the issue of how quanta behave in Shadows of Mind (Oxford: Oxford University Press, 1994), pp. 246-258.
2. Chaos theory based speculation concerning indeterministic neural processes can be found in Henry Stapp’s Mind Matter and Quantum Mechanics, (New York: Springer Verlag, 1993), George P. Scott’s "Dissipative Structures and the Mind-Body Problem," (in Time, Rhythms, and Chaos in the New Dialogue with Nature, ed. George P. Scott, Ames, IA: 1991, pp. 259-272), and Jesse Hobbs’ "Chaos and Indeterminism," (Canadian Journal of Philosophy 21: 141-164).
3. Robert Kane, The Significance of Free Will (SFW), (Oxford: Oxford University Press, 1996), pp. 32-35. Kane traces this principle back to Aristotle (Nichomachean Ethics, 1113b21, 1114a13-22, and 1114a18-19 and Physics, 255a8 and 256a6-8, both contained in The Works of Aristotle, ed. W.D. Ross, Oxford: Oxford University Press, 1915), Plotinus (Ennead 3.1.4 16-27, in The Philosophy of Plotinus, New York: Appleton, Century, Crofts, 1950), John Bramhall (The Works of John Bramhall, Oxford: John Henry Parker, 1844, p. 30 and p. 151), Thomas Reid (The Works of Thomas Reid, ed. William Hamilton, Hildeshein: George Ulm, 1983, p. 599), and Kant (Religion within the Bounds of Reason Alone, trans. T Greene and H. Hudson, New York: Harper and Row, 1960, p. 40). He cites Galen Strawson (Freedom and Belief, Oxford: Oxford University Press, 1986) Martha Klein (Determinism, Blameworthiness, and Deprivation, Oxford: Oxford University Press, 1990), as the contemporary philosophers who attach the most significance to the notion that free will entails being the ultimate source of one’s character. Other recent discussion of it are to be found in Paul Gomberg ("Free Will as Ultimate Responsibility," American Philosophical Quarterly 15, 1975: pp. 205-12), Thomas Nagel (The View from Nowhere, New York: Oxford University Press, 1986), Gary Watson ("Free Action and Free Will," Mind 96, 1987: 145-72), Ted Honderich (A Theory of Determinism, Oxford: Clarendon Press, 1988), Susan Wolf (Freedom within Reason, Oxford: Oxford University Press, 1990), James Lamb ("Evaluative Compatibilism and the Principle of Alternative Possibilities," Journal of Philosophy, 90, 1993: 517-27), David Widerker ("On an Argument for Incompatibilism," Analysis 47, 1987, pp. 37-41; "Frankfurt on ‘Ought’ Implies ‘Can’ and Alternative Possibilities," Analysis 49, 1991, pp. 222-4), Derk Pereboom ("Determinism Al Dente" Nous 29, March 1995, pp. 221-45; Living without Free Will, Cambridge: Cambridge University Press, 2001, pp. 4, 46-7), Saul Smilansky (Free Will and Illusion, Oxford: Oxford University Press, 2001, pp. 47-8, 52-5, 162-7, 209-12, 254), Bruce Waller (Freedom without Responsibility, Philadelphia: Temple University Press, 1990), and Richard Double (The Non-Reality of Free Will, Oxford: Oxford University Press, 1991).
4. I argue for this position in ________.
5. "Agent Causation as the Solution to all the Compatibilist’s Problems," (ACSCP) p. 1 (presented at the 2002 APA Pacific Division Meeting) and "A Compatibilist Version of the Theory of Agent Causation," (CVTAC) p. 273, Pacific Philosophical Quarterly 80 (1999) 257-77.
6. Cf. Timothy O’Connor, "Agent Causation," pp. 177-179; in Agents Causes and Events: Essays on Indeterminism and Free Will, ed. Timothy O’Connor, (Oxford: Oxford University Press, 1995), pp. 173-200.
7. ACSCP, pp. 1-5. In saying that an agent caused act is neither "indeterministic" nor has "any indeterminism in (its) causal history," Markosian appears to be denying both O’ Connor’s claim that an agent’s active power enables her "to freely and directly bring about (or not bring about) any of a range of effects" (O’ Connor, op. cit., p. 179; cf. also William Rowe "Indeterminism and Free Will," p. 160, Agents Causes and Events: Essays on Indeterminism and Free Will, pp. 151-171 and Thomas Reid, Essays on the Active Powers of Man, IV, ch. I, p. 512; in The Works of Thomas Reid DD, ed. Sir William Hamilton, James Thin, 1895) and Randolph Clark’s thesis that an agent caused act is only probabilisticly caused by its agent’s reasons. Cf. "Toward a Credible Agent-Causal Account of Free Will," pp. 203-204, in Agents Causes and Events: Essays on Indeterminism and Free Will, pp. 201-215.
8. Ibid., pp. 4-5.
9. Cf. O’Connor, op. cit., p. 186.
10. Markosian does not appeal to the mediate/immediate distinction in formulating his view; yet something like it applies there: in describing an example of agent causation, he says that "Yasmine’s saying ‘pass the salt please’ causes Imran’s passing (of) the salt in virtue of causing something else (namely, Imran’s causing himself to pass the salt)." (italics mine) ACSCP, p.4.
11. ACSCP, p. 3.
12. Cf. Lawrence Brian Lombard, Events: A Metaphysical Study (London: Routledge and Kegan Paul, 1986), pp. 111-114.
13. Cf. Nicholas Rescher, The Philosophy of Leibniz (New York: Prentice Hall, 1967), pp. 145-46 for a discussion of this doctrine. Here I am suggesting a modification of COMTAC so that if an agent performs a certain act as the (necessary) consequence of exercising her active power, then it is freely willed. Had she not exercised this power, ‘letting nature take its course’, and performed the action anyway, she would not have been responsible for it, as it would have been determined by events beyond her control. On the other hand, had she not performed the act, she would have been responsible for its not being done for having failed to agent cause its performance, not because she agent caused herself to refrain from performing it, which was beyond her power.
14. ACSCP, p. 5.
15. ACSCP p. 7.
16. Cf. Kane, op. cit, pp. 14-15, 60-78 and Smilansky, op. cit. pp. 40-41.
17. Cf. Kane, op. cit. pp. 32-35.
18. Cf. _______.
19. CVTAC pp. 269-271.
20. Kane, op. cit., pp. 122-123.
21. Peter van Inwagen, An Essay on Free Will, (Oxford: Oxford University Press, 1983), pp. 93ff.
22. John Martin Fischer and Eleonore Stump draw this distinction in "Transfer Principles and Moral Responsibility," Philosophical Perspectives 14, 2000, ed. James Tomberlin, p. 47.