A Christological Argument Against the Principle of Alternate Possibilities

Many (not all) advocates of libertarian free will endorse the Principle of Alternate Possibilities (PAP):

PAP: S is morally responsible for doing A only if S could have done otherwise.

PAP has come under continual fire ever since Harry Frankfurt’s seminal article in 1969, and many philosophers (including a number of leading libertarians) now accept that PAP is false. Leaving aside the philosophical arguments, however, it seems to me that any orthodox Christian ought to reject PAP on theological grounds.

Consider the following argument from the doctrine of Christ’s impeccability:

  1. Christ could not have sinned.
  2. Therefore, Christ could not have done otherwise than refrain from sinning.
  3. Christ was morally responsible when he refrained from sinning.
  4. If the ability to do otherwise is a necessary condition of moral responsibility, and Christ was morally responsible when he refrained from sinning, then Christ could have sinned.
  5. Therefore, the ability to do otherwise is not a necessary condition of moral responsibility (i.e., PAP is false).

The argument is deductively valid, so the advocate of PAP needs to reject one of the premises. It must come down to either 1 or 3, since 2 is just a restatement of 1, and 4 is a logical truth.

The defender of PAP might be most inclined to deny 1, but it must be recognized that there’s a very strong modal argument in favor of 1. If God is necessarily good then it’s logically impossible for God to sin; and if Christ is God Incarnate then it’s logically impossible for Christ to sin. Classical/Anselmian theism conjoined with Nicene/Chalcedonian Christology entails that Christ could not have sinned (in the strongest sense of “could not”). The logic of this argument is, one might say, impeccable.

So denying 1 isn’t an attractive option. But neither is denying 3, for several reasons. First, it entails that Christ’s sinless life was not morally praiseworthy, because moral responsibility is a necessary condition for moral praise. (What Christian would be comfortable saying that his life, despite all his failings, is more morally praiseworthy than Christ’s?)

Second, it seems to undermine the moral significance of Christ’s temptations. What does it matter that Christ was “one who in every respect has been tempted as we are, yet without sin” (Heb. 4:15) if Christ wasn’t morally responsible when he resisted temptation?

Third, it undermines the idea, widely held within the Christian tradition, that Christ merited our salvation by living a perfect life in obedience to the moral law. How could there be merit without moral responsibility?

I therefore conclude that anyone who holds to an orthodox Christology ought to reject PAP. Needless to say, this poses a problem for Arminians who want to wield PAP against their Calvinist brethren.

I very much doubt that I’m the first person to come up with this argument, but I don’t recall coming across it before.

23 Responses to A Christological Argument Against the Principle of Alternate Possibilities

  1. This is very good. You would likewise get a similar result if you applied Alvin Plantinga’s Transworld Depravity. You would be forced to conclude that in at least one of every conceivable world Christ “must” fail. Therefore, if Christ didn’t fail, then God was holding back, and the entire thing was a deterministic and dishonest sham. So premise 5): Therefore, the ability to do otherwise is not a necessary condition of moral responsibility (i.e., PAP is false), is valid!

  2. Dr. Anderson,

    Interesting observations. However I’ve never been too impressed by making God out to be necessarily morally perfect, as opposed to morally perfect in spite of having might not been so. This was never part of my old Christian belief, and I don’t see how it follows from any Biblical teaching, or even from any ancient tradition (as opposed to medieval developments, though I don’t mean to imply we should toss the latter aside lightly).

    Moreover your argument for (1) seems to require that Jesus not simply be God incarnate, but necessarily God incarnate. So even supposing God is necessarily morally perfect and Jesus is God incarnate, if Jesus might not have been God incarnate then we can’t immediately conclude that Jesus could not have sinned.

    On the other hand, it’s worth recognizing the motivation behind saying that Jesus could have sinned. Indeed we don’t want to just say that Jesus might have sinned, but rather that Jesus might have sinned while otherwise living a life identical to the one he did on the Christian view, up to any arbitrary time before his death. So in other words, it’s not enough for the naive PAP’er to propose that maybe Jesus could have been just another Jewish laborer in first-century Palestine. Rather, we want to say that he could have been a chosen prophet of God, spreading the Gospel, enduring his trials, but then at the last moment giving into temptation. So for example, it could have been the case that Jesus behaved exactly like he did in the New Testament, except that his last words be a bitter curse on God, his first and only sin at the moment just before death. Otherwise Jesus isn’t morally praiseworthy for all of his noble endurance on PAP.

    But if this is the case, it suggests that God the Father could have made a mistake in his Jesus-saves-the-earth plan (or his Jesus-saves-the-elect plan, if you prefer). But if the Father is essentially omniscient, then such mistakes would seem to be impossible. And while we might otherwise want to allow for the logical possibility of Jesus sinning, we are much more deeply committed to the essential omniscience of God. So at least by the occasion of Jesus’ baptism (at which time it is clear God has chosen Jesus), Jesus could not have sinned from that time forward. And so we must again give up PAP.

    Denying (2), of course, as you observe flies in the face of the fairly obvious fact that Jesus’ sinlessness is morally praiseworthy. However I’m not sure I buy your subsequent arguments for (2). It seems to me that the supposed necessity for Jesus’ sacrifice is quite mysterious whether or not one holds that Jesus was morally responsible for his sinlessness.

    Anyway, congrats on the clever argument.

    –B

    • Thanks for the comments, Ben.

      God’s necessary goodness isn’t just a philosophical intuition; it has direct biblical support, e.g., James 1:13. In fact, the argument for Christ’s impeccability could be premised with James 1:13.

      As for your second point, orthodoxy maintains that Christ was a divine person contingently conjoined with a human nature, not a divine person contingently conjoined with a human person (the heresy of Nestorianism). Assuming (as I do) transworld identity, there’s no possible world in which Jesus exists and is not God Incarnate. There might be a first-century Jew who looks like Jesus, speaks like Jesus, etc., but that person would not be Jesus Christ.

  3. Excuse me, I meant “(3),” not “(2).”

  4. Hi James,

    I’m generally disposed to agree, but: it does seem like denying (1) has a tradition (a small one, to be sure) behind it—and a Reformed pedigree to boot! C. Hodge denied (1), opting for the sinlessness view over against the impeccability view. More broadly, Millard Erickson has denied (1), as well as several other contemporary theologians. Tradition is on the side of (1), and I agree with (1), but in response to your argument against (1) they would argue that you’re not taking into account what must be the case for Christ to be fully human. As Hodge says, “If He was a true man He must have been capable of sinning. . . .Temptation implies the possibility of sin. If from the constitution of his person it was impossible for Christ to sin, then his temptation was unreal and without effect, and He cannot sympathize with his people.”

    In any event, this would be a tough position to hold, I think. But, to press further, one might perhaps argue that ability to sin is a particular property of the second person of the Trinity, and not the other two persons, and so not essentially belonging to the divine essence. And, some might give up necessary goodness, saying that it makes God praiseworthy that he could sin but doesn’t. It’s a display of his power and self-control. And some might modus tollens the argument: PAP is not false, thus Jesus could sin. Again, I don’t think these ultimately will work, but I think someone might be able to give (1) a better go at it than immediately falling to the necessity objection. But I think Crisp makes all the right moves against the sinlessness view.

    But having said that, here’s another, more pressing worry: PAP (and there’s many definitions of this term, some of which might affect your argument more than others) doesn’t specify that if one is morally responsible for A-ing, where A is a good action, that one must be able to do some other action, B, where B is a *bad* (or sinful) action. It simply states one must be able to do otherwise than what one did. But it seems to me that PAP is *neutral* with respect to what other actions the agent must be able to do. For your argument to go through, it seems, you need to argue that Christ could not do otherwise than what he did, period. For your definition of PAP doesn’t include some rider about ‘morally *opposite* actions.’

    Thoughts?

    • Paul,

      I’m generally disposed to agree…

      Gee, thanks for the ringing endorsement. ;)

      I recognize that some of the Reformed have denied Christ’s impeccability, but like Crisp, I think the denial is poorly motivated. (Actually, this post was inspired by the relevant chapter in Crisp’s God Incarnate.) The arguments for (1) are stronger than the arguments against, at least for a classical theist.

      As for the “more pressing worry”: I guess I don’t see the problem. My statement of PAP is the standard one. If the relevant action (or intentional behavior) is “refraining from sinning” then “doing otherwise” is simply “sinning”. Yes, there could be multiple ways in which one could do otherwise than refraining from sinning — but that’s beside the point here.

  5. James,

    I agree with the first point.

    On the second, why think ‘refraining from sinning’ is an *action*? Aside from that, doing *otherwise* doesn’t mean “morally opposite.” It means, refraining from doing that action. Doing *something* else. PAP does not say that one is morally responsible for doing a good action only if one could have done a bad action instead. PAP’s just neutral on that, but your argument seems to depend on that read. (Crips makes this point in the relevant chapter vis-a-vis God. It’s impossible for God to sin, he must always do right, but there’s many right actions he could do, so God could “do otherwise” than he does.) Moreover, there’s various ways of cashing out PAP. Some cash it out as simply “refraining from doing A.” Finally, another ‘libertarian’, Wolf or Ekstrom (forgot right now) denies PAP is needed for morally praiseworthy actions, just for blameworthy ones. So that’s the worry. Not sure if I motivated it any further?

  6. James, sorry, I typed the above pretty much while running out the door to go to church this morning! Let me state “the more pressing worry” this way:

    We have:

    PAP: S is morally responsible for doing A only if S could have done otherwise.

    It seems that your read of this is that the ‘otherwise’ has to be an action of opposite moral value. If this were true, then no Christian should hold to PAP if they believe God is necessarily good. For it is impossible for God to sin, so it is impossible for God to do other than good or sinless actions. But libertarians who ascribe LFW (with PAP) to God will say that God has alternate possibilities. God could do otherwise than he did, yet none of the “other” things he could have done are wrong or sinful things.

    If the ‘otherwise’ in PAP is neutral with respect to the value of the actions the relevant agent may do, I don’t see how only being able to do good actions means PAP isn’t in play. So it seems you are adding qualifiers to PAP, such that:

    PAP*: S is morally responsible for doing A, where A is a good action, only if S could have done otherwise than A, ¬A, where ¬A is a bad action, and vice versa.

    But I don’t see that assumption among libertarians. From Mckenna and Widerker’s Moral Responsibility and Alternative Possibilities, here’s some statements on PAP:

    “PAP: An agent is morally responsible for performing a given action A only if he could have avoided performing it” (Widerker, p.53).

    “PAP: An agent S is morally responsible for its being the case that p only if S could have made it not the case that p.” (Ginet, 75).

    Since there seem to be dozens, if not hundreds, of different ways Jesus could have avoiding performing some morally responsible action of his and none of those ways be *sinful*, then he satisfies PAP. Granted, Jesus could not have sinned, and so he was not free to sin, but I don’t see how that rules out PAP *unless* for PAP to hold the ‘otherwise’ is cashed out in terms of doing another action of opposite moral worth, and not just doing another action.

    Not sure if this was any clearer! :-)

    • Paul,

      It seems that your read of this is that the ‘otherwise’ has to be an action of opposite moral value.

      No, not at all. I’m only reading it such that ‘otherwise’ means ‘otherwise’. :) Doing otherwise than refraining from sinning is… sinning.

      If this were true, then no Christian should hold to PAP if they believe God is necessarily good.

      Actually, I think that too. :)

      But it seems to me that the Christological argument has more bite because Christ refrained from sin as a man. So one can’t pull “God’s not like us” card.

      It may be that part of our disagreement is that you’re viewing “doing A” in terms of action-tokens rather than action-types. If so, I can explain why I don’t think that’s the right way to read PAP. If “doing A” is understood as an action-type, then “refraining from sinning” is a proper instance of “doing A”.

      Perhaps a concrete example will help. Jesus is tempted by Satan to worship him (Matt. 4). Jesus can either worship Satan or refrain from worshiping Satan (and in either case, in any number of ways). But according to PAP, if Jesus is morally responsible for refraining from worshiping Satan, the relevant ‘otherwise’ here is simply worshiping Satan. Right? What’s relevant here is the action-type, not the specific action-token.

  7. James,

    I appreciate the action-type/token distinction. I thought that’s what was going on too, and am glad you confirmed it. I do think that if your argument goes through, it’s a powerful one for the Christian. I just don’t think it does, let me explain briefly why:

    First, there’s an ambiguity in the basically Frankfurtian definition of PAP you gave. Sometimes, especially earlier on, ‘responsible’ is not always properly distinguished from ‘blameworthy.’ Thus, as Haji notes in *Deontic Morality and Control*, the proper characterization of PAP is something like:

    PAP*: S is *blameworthy* for A only if S could have refrained from A-ing.

    This would change the dialectic of your argument, for you’d need to find an action Jesus was *blameworthy* for and which he could have done otherwise. The problem is that, as your argument points out, there are no such actions! :-)

    David Copp makes this same point in his paper: “Ought Implies Can, Blameworthiness, and the Principle of Alternative Possibilities.” Copp interprets Frankfurt as not always being clear here, and sometimes meaning to refer to blameworthiness than responsibility in general, but this is a vexed hermeneutical question. It should be said that *blameworthiness* is the relevant term in the debate over the Widerker defense, or, the W-defense. This is the context of the Widerker-Mckenna debate. So I think it’s fair to point to the ambiguity.

    At any rate, there’s something important here. For it’s not clear to me that libertarianism is committed or dependent on the idea that PAP applies to morally responsible actions for which we are *praiseworthy.* Indeed, Susan Wolf argues that PAP only applies to blameworthiness and not praiseworthiness. So this is one way the PAPist could respond. (And I also put aside definitions of PAP which seem to get around your objection, e.g., Shabo: “PAPm1: an agent is morally responsible for her action only if it was possible that she perform that action under a different mode.)

    Second, aside from the above, I’m still not 100% sold on your argument (though I *want* to be, and I *used* to be :-). You’re casting matters negatively, i.e., Jesus’ *refraining* from doing something. At a minimum, it’s contentious whether refraining from acting is an action. There’s debate here. Moreover, some would say that if you can’t do X, you can’t refrain from doing X (e.g., Kent Bach, “Refraining, Omitting, and Negative Action,” *A Companion to the Philosophy of Action*). (And we should note that you need ‘refraining from refraining’ if A in your PAP definition is cashed out as *refraining* from sinning. And I’m really not sure if refraining from refraining is an action!) Moreover, Fischer and Ravizza claim that one can’t be responsible for *omissions* if one couldn’t have done otherwise, though Frankfurt demurs (though both seem to distinguish omissions from actions).

    But even if we put this aside, there’s *still* PAPs Jesus could have, even on your token-type distinction. Here’s how you put it:

    But according to PAP, if Jesus is morally responsible for refraining from worshiping Satan, the relevant ‘otherwise’ here is simply worshiping Satan. Right? What’s relevant here is the action-type, not the specific action-token.

    So, on your view if Jesus is responsible for refraining from worshipping Satan, then Jesus must have been able to do otherwise, where this means, not refraining from worshiping Satan. Well, Jesus could have destroyed Satan. Jesus could have made Satan immediately withdraw the offer. These don’t seem identical to merely “not refraining from worshiping Satan.” That is, Jesus could say, “I didn’t merely ‘not refrain’ from worshiping Satan, I turned him into a poached egg!”

    Lastly, it’s not the case that PAPists affirm PAP for all actions across the board. They allow that PAP could be false where will-setting has taken place. They would require PAP at the level of will-setting, though (this is where tracing comes into play). But Jesus’ will was already ‘set’ from the start. Thus you should turn your attention to whether Jesus undermines libertarian requirements for will-setting. But this doesn’t seem too promising. First, paradigmatic defenders of will-setting constraints, e.g., Robert Kane, *don’t* apply that constraint to God, who has “self-perfecting” freedom (I bypass the issue of whether God’s inability to sin refutes PAP, which I don’t think it does; and, if it does, it only refutes it as applied to *any* free creature, rather than *some*). Second, it seems Jesus will with respect to his human nature was already set. He didn’t ‘form’ it in the way us ordinary sinners do (assuming the libertarian story here). But, even if Jesus did form or set his human nature, he would have done so at a very young age, and may have had a variety of “alternative *good* possibilities” to choose from. This wouldn’t be revealed, and so we couldn’t say that when he set his will it was in a situation in which he had either good or bad choices. Since you admitted that ‘otherwise’ doesn’t need to be ‘morally opposite,’ and since we don’t know what actions Jesus would have done to ‘set’ his will, we (you) can’t say that the “otherwises” he had available to him were morally bad ones.

    Thus, the libertarian could say that Jesus set his will at t and that tt, because we can trace it back to when Jesus libertarianly set his will at t. At t, though, Jesus had alternative possibilities, none of which were morally bad. If you argue that PAP *does* commit one to claiming that the ‘otherwises’ must be morally opposite, this (i) cuts against your specific denial that you were committed to this, and (ii) doesn’t seem to be what the libertarian was ever committed to. So it does seem you must be committed to the “morally opposite” read. I don’t think that’s a proper read, though. For libertarians have explicitly denied it. Anyway, those are my reasons, such as they are :-)

  8. Sorry for the oddness in the last paragraph! I put greater than/less than symbols, and I guess it read me as trying to do html code. The point was Jesus set his (human) will at t, and so at times greater than t, there’s no APs, but there were APs at t, and t is less than all the revealed actions of Jesus, thus you can’t use any of them as an example against PAP. So you need to appeal to the APs present at t, and perhaps those who all good, unless you think PAP implies moral opposites . . .

  9. Paul,

    As usual, you raise some very good points. Briefly:

    1. If ‘responsible’ is interpreted merely as ‘blameworthy’, then yes, all bets are off. But it’s odd to me, given the basic symmetry between praiseworthiness and blameworthiness, that the AP condition would apply only to the latter.

    2. If “refraining from sinning” isn’t properly an action, just replace it with “resisting the temptation to sin” or however you want to describe Jesus’ positive response to the temptations of Satan (as a general action-type).

    3. Regarding will-setting, I’m not sure this is a good out for the PAPist (for want of a better label!). Presumably Jesus was impeccable from birth (even conception) in which case there was no will-setting process. Certainly there was no will-setting process involving prior choices featuring morally significant APs (which is the usual idea). If anything, we also have a Christological argument against the will-setting condition.

  10. Hi James, thanks.

    On 1: Well, the asymmetry is discussed in the literature, and theologians have discussed it vis-a-vis God’s praisworthiness apart from his ability to do evil, but I’ll assume further discussion here would be a digression as well as that we both have other priorities and time constraints right now.

    On 2: That’s fine, but I think you can still get APs here—there’s more than one possible way to resist sin. (And I’m not sure action types are the right taxonomy, I think ‘otherwise’ means doing something other than he *did*, either some other action, or refraining from acting.)

    On 3. Sorry if I wasn’t more clear. I think your argument is best taken as an argument against libertarian accounts of *origination*, or source in/compatibilism. The easy escape from the argument is to claim Jesus was acting out of a formed or set character, and PAP isn’t needed—just like it isn’t for all of our actions. So, the libertarian will have to argue either (a) Jesus didn’t have to form his will or (b) Jesus did form his (human) will. Now, I’m not 100% sure that (b) is false; at least, it’s open for me that the libertarian could make this move. Jesus formed his human will, but could only form it to be morally virtuous (and he had APs in the process). But, (b) may be false (maybe you have an argument that it is?). On the other hand, if the libertarian opts for (a), he’ll have to show some relevant disanalogy between Jesus and us. If he can do so while maintaining Jesus full humanity and his being “like us in every way,” then he’d have an out. It’s not fully clear to me that in order for Jesus to be like us in every way his will must be set in the way libertarians think *sinful* men must set their wills. But I’ll leave it to the libertarian to argue all of this.

    The upshot is, I think your argument is better cast as an argument against source incompatibilism (with interesting implications for Galen Strawson’s argument against responsibility, as well as some very interesting implications for ‘manipulation arguments,’ both of which focus on origination of character instead of PAPs, though the latter’s always in the mix somehow/where too).

    • Paul,

      I confess I don’t understand your “On 2″ and “On 3″, but I’m happy for you to have the last word. :)

  11. A quick note on (1): I just received Dana Nelkin’s *Making Sense of Freedom and Responsibility* and so I peeked through it. She argues for a compatibilist account of agent causation and OIC. Anyway, she has a helpful explanation for why *blameworthiness* has been thought to be the relevant ascription in PAP, and even argues that this is what Frankfurt believed, though he conflated responsibility and blameworthiness. And the answer for why lies in the OIC principle. Basically, it is not the case that for every x, if x is a praiseworthy action, S ought to have X-ed. But, for every y, if y is a blameworthy action, S ought not have Y-ed; thus S could have done otherwise. That’s the quick and dirty of it, in any event.

    • Okay, not quite the last word. :)

      I agree that as a general principle “it is not the case that for every x, if x is a praiseworthy action, S ought to have X-ed.” But I think it’s true for the relevant Christological subset (obedience to the moral law, resisting the temptations of Satan, etc.). At any rate, there is a positive moral counterpart to blameworthiness, call it what you will, that falls under the category of moral responsibility.

  12. James, sorry for the confusion:

    On 2: There’s still APs even with your restatement of the action.

    On 3: Your argument in this blog post is better cast as an argument against source incompatibilism.

    Hopefully these are now understandable (now that I’ve removed all the cause for confusion! :-)

    On your not quite last word: Maybe there is a positive counterpart to blameworthiness, but when you turn to talk about what Jesus *ought* to do the libertarian will simply say that Jesus *can* do those morally obligatory praiseworthy actions (and clearly Jesus can do them!). But why think that *this* requires him to have been able to do “otherwise?” So if we construe PAP as referring to blame, this gets around your argument. Moreover, the PAPist is not committed to saying that if S does a *praiseworthy* action, S must have been able to do otherwise, even if this action is something S ought to have done. The PAPist will just say that if S ought to have done A, then S can do A. So I think the worry’s still there.

  13. I think Paul did an excellent job showing why the argument doesn’t work. But, since there still seems to be some hesitance to rejecting it, I thought I might add a few points to help.

    1. Christ could not have sinned.
    2. Therefore, Christ could not have done otherwise than refrain from sinning.

    My contention is with P2, the sub-conclusion, and, by extension, P3, which are false by definition. For beginning with P2 the argument equivocates on the meaning of “refrain.”

    If some act x is a logical or metaphysical impossibility for some agent P then there is no substantive or meaningful sense in which P could refrain from x. For one cannot refrain from an act unless, as Paul put it, refraining from refraining is logically possible; that is to say, the negation of refraining is a genuine possibility; or, more simply, it is the case that P’s will is involved in whether P can or cannot do x.

    Consider the following:
    Square-circles are logically impossible entities: not even God can create one. In what meaningful sense, then, can God be said to will to refrain from creating square-circles? There is no meaningful sense in which God, or any person, can be said to refrain from creating square-circles. The point being that the CAAPAP only works if logical contradictions are logically or metaphysically possible. But this is just to speak nonsense.

    Thus, Paul was correct when he said the PAPist (this label is amusing, by the way) need only believe that Christ could have refrained from sinning in some other morally perfect way than he did when he actually refrained from sinning in order for the PAP to hold. To formalize this response:

    1. Christ could not have sinned.
    2. Therefore, Christ could not have done otherwise than refrain from sinning.
    3. Christ was morally responsible when he refrained from sinning.
    4. If the ability to do otherwise is a necessary condition of moral responsibility, and Christ was morally responsible when he refrained from sinning, then it is necessary that Christ could have refrained in some other morally perfect way than he actually did.
    5. Christ could have refrained from sinning in some other morally perfect way than he actually did.
    6. Therefore, Christ was libertarianly free and the PAP holds.

    As you may have noticed, if we are going to speak in action-types, rather than action-tokens, then you better be able to provide a plausible argument for denying that the set of action-types for a particular circumstance is no greater than one token. All you have done here in the CAAPAP is make the assumption that the action-type of “refraining from sin” refers to only one possible action-token, but that is just begging the question against the PAP.

  14. Correcting a typo:
    If we are going to speak in action-types, rather than action-tokens, then you better be able to provide a plausible argument for denying that the set of action-tokens of some action-type regarding a particular circumstance is no greater than one token.

  15. It occurred to me that I was careless and forgot to make it explicit exactly how I was using the term “refrain” in my reformulation. So I thought it would be beneficial to provide some formal definitions for how the term has been used. This should be especially useful for understanding my reasoning in regards to when I claimed that the meaning of “refrain” was being equivocated in the original argument.

    Definitions for the term “refrain”:

    (D1) P can not do some act x.

    (D2) P does some act y and y is not some act x.

    (D3) A label or reference to a set of action-tokens.

    (D1) P can not do some act x.
    D1 is, strictly speaking, the correct definition for “refrain.” And it is very important to understand that the expression “can not” is NOT identical to “cannot.” This is particularly significant when considering P2 of the CAAPAP. For P1 claims that some person C cannot x, which necessarily implies that the claim, C can x, is false. While P2 concludes from P1 that C *can not* x. However, this is a non sequitur because the claim “C can not x” implies that C can x but merely does not x, which contradicts P1. Therefore, if D1 is the intended meaning of “refrain” then the argument is invalid.

    (D2) P does some act y and y is not some act x.
    D2 might be a more colloquial usage of the term but, in my opinion, it should be shunned. This is where the ambiguity sets in, since it is not entirely clear whether D1 or D2 was intended in the CAAPAP. The primary distinction between these two definitions being that D2 makes no implication concerning whether C can x. For, in context to the CAAPAP, this would make P2 synonymous to “Therefore, Christ could not have done otherwise than some morally good act that is not a sinful act.” But, as you will notice, this is essentially a restatement of P1, or an identity claim. Thus, P1-P2 is really saying nothing more than “A. Therefore, A.” While this is not particularly interesting, this does present some other difficulties in the argument. For example, on this interpretation P4 would be read, “If the ability to do otherwise is a necessary condition of moral responsibility, and Christ was morally responsible when he did some morally good act and not a sinful act, then Christ could have sinned.” And even though this may be a valid argument if we are using the material implication or conditional, I highly doubt the material conditional was intended. For there is obviously no relevant relation between Christ doing some morally good and whether he was able to sin. Therefore, I suspect the relevant implication is intended in P4. But, once this is made evident, the argument makes no sense. Therefore, D2 cannot be consistently used as the intended meaning for “refrain.”

    However, D2 is also the way I used the expression in my reformulation. My apologies for not stating this explicitly. I had it in mind to do so but forgot to. So at the expense of being redundant, this is what I should have said:

    1. Christ could not have sinned.
    2. Therefore, Christ could not have done otherwise than some morally good act.
    3. Christ was morally responsible when he did some morally good act.
    4. If the ability to do otherwise is a necessary condition of moral responsibility, and Christ was morally responsible when he did some morally good act, then it is necessary that Christ could have done some other morally good act than what he actually did.
    5. Christ could have done some other morally good act than what he actually did.
    6. Therefore, Christ was libertarianly free and the PAP holds.

    (D3) A label or reference to a set of action-tokens.
    And, lastly, this third usage is a construct that arose from the subsequent conversation about the argument, that is, from the action-type/token distinction. On this interpretation the expression “refrain from sin” is merely a reference to a set of objects (e.g., events or acts). In other words, we might say there is, for example, an action-type such that it is the set of action-tokens {x, y, z}. This is obviously not a typical way to define “refrain.” Nevertheless, if D3 is the intended meaning of “refrain” then, as I said previously, to assume that the set only contains one action-token without some reason for supposing this is the case is to beg the question against the PAP, which requires a minimum of two action-tokens.

  16. Jordan,

    Thanks for the comments. But as I pointed out in my reply to Paul here, the argument can easily be reformulated without the term ‘refrain’.

  17. Pingback: Friday | favorites | cheap books | giveaway | gluttony | Kirk Cameron