[1] Advocates of libertarian freedom have often affirmed the principle of alternate possibilities (PAP): A person is morally responsible for what he has done only if he could have done otherwise. As some see it moral responsibility and divine foreknowledge are mutually exclusive.
[2] If this argument were sound and valid none of us could ever be praiseworthy or blameworthy. This runs counter to both experience and Scripture. So, at least one of the premises must be false. Which one(s)? Open Theists deny 3. Calvinists, advocates of compatibilist freedom, deny 2. Molinists deny 1. All of these responses have received a good deal of discussion. There is, however, another option, that has not been given nearly as much attention.
[3] In an article surveying recent work on divine foreknowledge and human freedom, Linda Zagzebski refers to the denial of both determinism and the principle of alternate possibilities as ‘the newest solution in the literature.’ [1] The idea here is that one can consistently affirm libertarian freedom, in which genuine freedom is incompatible with causal determinism, and deny the principle of alternate possibilities (PAP): A person is morally responsible for what he has done only if he could have done otherwise. Taking our cue from John Martin Fischer let’s call this view ‘hyper-incompabtibilism.’ [2]
[4] Advocates of hyper-incompatibilism appeal to so-called Frankfurt counterexamples, thought experiments designed to show that neither genuine freedom nor moral responsibility requires alternate possibilities. Consider for example, William Lane Craig’s comments:
[5] In a discussion of Augustine’s response to the alleged incompatibility of divine foreknowledge and human freedom, David Hunt has the following to say:
[6] With the comments of Hunt and Craig in mind we can return to the original argument:
[7] If hyper-incompatibilism were correct, libertarian Arminians could join Calvinist compatiblists in the denial of premise 2. However, before we celebrate this family reunion we had better have a closer look at hyper-incompatibilism. In order to do so, it will be helpful to have a Frankfurt case for reference.
[8] Suppose Satan is monitoring Adam’s brain. If Satan sees that Adam will obey God, Satan shocks Adam such that Adam decides to disobey. If Satan sees that Adam will disobey without any interference, Satan does nothing. As a matter of fact, Satan is aware that Adam will disobey without any demonic assistance, and Adam subsequently does so.
[9] If Craig and Hunt are correct, Adam disobeyed freely, because there was not any Satanic interference, and so Adam is responsible for his act, even though he could not have done otherwise. To see whether Hunt and Craig are right about this let’s think some more about this case.
[10] Suppose Adam’s choice was indeed made in the absence of alternative possibilities. That is to say, suppose that at the moment prior to choosing, Adam’s future was not one of forking garden paths but one in which there was but one and only one possible future, choosing to disobey God. For convenience let’s call the state of affairs Adam was in when he chose, Eden. If in Eden, Adam cannot but choose to disobey, then it seems that Adam’s situation is accurately described by the de dicto necessity, ‘Necessarily if in Eden, Adam disobeys God’ and the equivalent de re necessity, ‘Adam is necessarily such that, in Eden, Adam disobeys God.’ One who denies that Adam’s situation is rightly described by this pair of claims would need to allow that, in Eden, Adam could have done other than disobey God. That is to say, one would have to deny that Adam’s situation is a Frankfurt case.
[11] If Adam’s circumstance is indeed Frankfurtian then Adam’s absence of alternatives-the fact that he could not but choose to disobey God-does not rule out Adam’s responsibility. However, Adam cannot be morally responsible for disobeying in Eden, for doing so is one of his essential properties. Apart from embracing Cartesian voluntarism, no one, not even God, can be responsible for one of her or his essential properties. And, as Leibniz rightly noted, the Cartesian view carries a very high price tag, the rejection of the law of non-contradiction:
[12] Assuming that libertarians do not want to join the ranks of Cartesian voluntarists, there is no reason to think that Frankfurt cases are counterexamples to the principle of alternate possibilities (PAP): A person is morally responsible for what he has done only if he could have done otherwise.
[13] Reflection on Adam’s situation allows us to develop a procedure for responding to Frankfurtian claims that alternate possibilities are not a necessary condition for moral responsibility. Take the state of affairs, S, in which there is but one possible action or event, E, and for which an agent, A, is said to be responsible. Point out that, given an essential feature of Frankfurt cases, that is, the absence of alternative actions, the de dicto necessity ‘Necessarily, if in S, A brings about E’ and the corresponding de re necessity, ‘A is necessarily such that, if in S, A brings about E’ are true. Note that if hyper-incompatibilists were right in thinking that having alternative possibilities is not a necessary condition for moral responsibility, one could not rule out the possibility of A’s responsibility for an essential property. But, it is absurd to suppose that A could be responsible for one of A’s essential properties. Hence, hyper-incompatibilism is not a viable option.
[14] One might wonder about the soundness of my approach. It may appear that my strategy for refuting Frankfurt cases succeeds only if one is willing to reject moral responsibility even in cases in which an agent could have done otherwise. To see why consider the call of Abraham. Suppose that due to a Satanic oversight, when God commands Abraham to leave his homeland, Abraham has alternate possibilities, and makes the hard choice to leave his homeland.
[15] Let ‘S’ stand for the situation Abraham is in immediately prior to his decision. Since Abraham has libertarian freedom in S, there is a possible world, Beta, in which subsequent to S, Abraham decides to stay put, and a possible world, Kappa, in which subsequent to S, Abraham decides to leave for the promised land. Allowing that possible worlds do not possess their constituent states of affairs contingently, it seems to follow that if Abraham is in S in Beta then it is necessarily the case that Abraham decides to stay. And so we have the de dicto necessity, ‘Necessarily if Abraham is in S in Beta then he decides to stay.’ And, similarly if Abraham is in S in Kappa it is necessarily the case that he decides to leave, that is, ‘Necessarily if Abraham is in S in Kappa he decides to leave.’
[16] For the de dicto necessity, ‘Necessarily if Abraham is in S in Beta then he decides to stay’ there is a corresponding de re necessity, ‘Abraham is necessarily such that if in S in Beta he decides to stay.’ So too for the de dicto necessity, ‘Necessarily if Abraham is in S in Kappa he decides to leave’ there is a corresponding de re necessity, ‘Abraham is necessarily such that if in S in Kappa he decides to leave.’ But, according to my strategy for refuting Frankfurt cases, it is absurd to suppose that Abraham could be responsible for one of his essential properties. It seems then to follow that, if successful, my strategy would rule out moral responsibility even in cases in which an agent could have done otherwise. Call this objection the no-libertarians-allowed objection.
[17] Not wishing to preclude the possibility of moral responsibility in cases in which an agent could have done otherwise, I would be worried if the no-libertarians-allowed objection were foolproof. In fact I think that this criticism misses the mark. To understand why consider the relevant de dicto claims, ‘Necessarily if Abraham is in S in Beta then he decides to stay’ and , ‘Necessarily if Abraham is in S in Kappa he decides to leave.’ Note that world Beta differs from world Kappa in that in the former the state of affairs S is followed by Abraham’s deciding to stay, while in the latter the state of affairs S is followed by Abraham’s deciding to leave. In other words, Beta and Kappa branch after S. Hence, S in Beta is identical to S in Kappa.
[18] If S in Beta is identical to S in Kappa then any state of affairs compatible with S in Beta is compatible with S in Kappa and vice versa. Since S in Beta is compatible with Abraham’s deciding to stay, S in Kappa is compatible with Abraham’s deciding to stay. And, similarly since S in Kappa is compatible with Abraham’s deciding to leave, so S in Beta is compatible with Abraham’s deciding to leave. But if S in Beta is compatible with Abraham’s deciding to leave, then the de dicto claim ‘Necessarily if Abraham is in S in Beta then he decides stay’ along with its de re equivalent, ‘Abraham is necessarily such that if in S in Beta then Abraham decides to stay’ is false. And if S in Kappa is compatible with Abraham’s deciding to stay then ‘Necessarily if Abraham is in S in Kappa he decides leave’ and ‘Abraham is necessarily such that if in S in Kappa he decides to leave’ are false. It follows then that Abraham has neither deciding to stay if in S in Beta nor deciding to leave if in S in Kappa as an essential property. In sum, the no-libertarians-allowed objection errs by failing to see that since world Beta and world Kappa branch after S, S in Beta and S in Kappa rigidly designate the same state of affairs.
[19] Though the criticism misses the mark, the no-libertarians-allowed objection does focus on a key point; I do assume that if an agent A’s action E is rightly viewed as one of A’s essential properties then A cannot be responsible for that action. One whose modal intuitions run along the lines of David Lewis’s counterpart theory might disagree. On a counterpart theory even if it were the case that A could not possibly lack the property of bringing about E in S, it would not necessarily follow that this is one of A’s essential properties. So long as there is another possible world with an individual very much like A, a counterpart of A, who does not possess the property of bringing about E in S, that property would be rightly viewed as one of A’s accidental properties. But, then, contrary to my objection, it appears that one could allow that the truth of the aforementioned necessities need not preclude the possibility of A’s moral responsibility for bringing about E in S. Call this objection the counterpart objection.
[20] Suppose, in accordance with the objection, that Abraham has a counterpart and Abraham’s counterpart does not possess the property of choosing to leave in S. Assume as well, that it follows from this that the property of choosing to leave in S is one of Abraham’s accidental properties. If so, then one can indeed appeal to counterpart theory to show that Frankfurt cases do not lead to the absurdity that it could be an open question whether or not an agent is responsible for some of her or his essential properties.
[21] Though appeal to counterpart theory may seem to save the Frankfurt cases from my critique, this is not so. The heart of my objection is that it is absurd to suppose that we cannot rule out the possibility of an agent’s moral responsibility for an action, when so acting is a property the agent could not possibly lack. I am willing to make the counterpart theorist a present of the term, ‘accidental property,’ such that one can have accidental properties that one could not possibly lack. I’ll then simply rephrase my objection such that instead of talking of essential properties, I’ll talk about properties one could not possibly lack. Calling some of these properties ‘accidental properties’ does not rescue the Frankfurt cases from my objection.
[22] One need not be a counterpart enthusiast to worry about the soundness of my critique. One might argue that it rests upon a failure to distinguish between ‘A is necessarily such that, if in S, A brings about E’ and ‘A does E in S.’ According to this objection I have correctly argued that A could not be responsible for the fact that ‘A is necessarily such that, if in S, A brings about E’ but that is irrelevant. One who appeals to a Frankfurt case is not concerned with questions regarding an agent’s responsibility for the fact that the agent could not have chosen otherwise (‘A is necessarily such that, if in S, A brings about E’). Rather, one who appeals to a Frankfurt case does so to show that acting in a state of affairs in which an agent could not have chosen otherwise (cf. ‘A does E in S’), does not preclude moral responsibility. While allowing that an agent A might be responsible for the fact that ‘A is necessarily such that, if in S, A brings about E’ might lead to absurd results, no such absurdities follow from allowing that A might be responsible for ‘doing E in S.’ Call this the irrelevance objection.
[23] By way of response, consider that which the objector takes to be the relevant state of affairs, i.e. A’s doing E in S. Since the state of affairs, S, is a Frankfurt state of affairs, to say that ‘A does E in S’ is to say that ‘In S, A does E, and could not have done otherwise.’ But then ‘In S, A does E, and could not have done otherwise’ is true, if and only if ‘A is necessarily such that, if in S, A brings about E’ is true. Therefore I have not erred by focusing my discussion of moral responsibility on ‘A is necessarily such that, if in S, A brings about E.’
[24] As a final objection, some might suggest that my discussion is fundamentally flawed in that my focus is on logical necessity, when it ought to have been on natural, i.e. causal necessity. According to these folks the relevant connections are natural not logical. After all, Frankfurt cases concern the possibility of interference with the brain, hence the frequent mention of neurosurgeons.
[25] The idea here is that when a Frankfurt advocate claims ‘Adam cannot but choose to disobey’ she is not asserting a logical necessity, but a natural necessity. ‘Adam cannot but choose to disobey’ is no more a logical necessity than ‘Adam cannot thrive on a diet of day-old donuts, and double shots of espresso.’ While it may be true in our world, with the dietary needs that we happen to have, that Adam won’t do well living on stale donuts and espresso, it is not logically impossible that Adam thrive on such a diet. The denial of ‘Adam cannot thrive on a diet of day-old donuts, and double shots of espresso’ is not a contradiction, and so ‘Adam cannot thrive…’ it is not a logical necessity, that is, a de dicto necessity with a corresponding de re necessity. And the same goes for ‘Adam cannot but choose to disobey.’ To maintain otherwise is to confuse causal and logical necessity. Call this objection the causal objection.
[26] Insofar as the causal objector wishes to underscore the fact that the state of affairs that Adam is in, S, and in which he cannot but choose to disobey, is a contingent state of affairs, I have no complaints. Claims like ‘Adam exists,’ ‘Adam chooses to disobey’ and the like, are, if true, contingently true. But the same cannot be said for the conditional claim, ‘ If in S Adam chooses to disobey.’ This claim would be contingently true only if there were two possible worlds that are identical up to and including S, but differ such that in one of the worlds Adam chooses to disobey, and in the other Adam chooses to obey. However to allow that there are two such worlds is to allow that Adam could have done other than disobey. But then Adam’s scenario no longer qualifies as a Frankfurt case. Given that the scenario is a Frankfurt case, it follows that there are not any possible worlds in which Adam does other than choose to disobey, in S. Hence, the conditional, ‘If, in S, Adam chooses to disobey’ is a de dicto necessity and there is a corresponding de re necessity, ‘Adam is necessarily such that, in S, Adam chooses to disobey.’
[27] In sum, William Lane Craig and David Hunt have appealed to Frankfurt cases in support of the claim that libertarian freedom and moral responsibility do not presuppose the truth of PAP: a person is morally responsible only if they could have done otherwise. I’ve argued that this so-called hyper-incompatibilist perspective has an absurd implication: one could be responsible for some of one’s essential properties. As I see it reflection on the implications of Frankfurt cases only increases the plausibility of PAP. [7]