What's the Problem With Penal Substitution?

Isn't the problem that the argument assumes that the demands of justice are agent-neutral? But the demands are not agent-neutral, are they? If Smith murders Jones, does justice demand that (1), (2) or (3)?

(1) Some punishment is suffered for the injustice.

(2) Smith suffers some punishment for the injustice.

(3) Smith or some fully informed, rational and freely consenting person P suffers the punishment.

It is definitely not (1), since we can't (morally) punish just anyone for Smith's behavior. (3) seems true only in cases where Smith's debt is transferrable. For instance, if P pays Smith's fine, it does seem that Smith is now indebted to P (P can just pardon the debt). But it does not seem like all debts are transferrable. If P serves Smith's jail term, does Smith owe P time in jail? I'm not sure what that could mean. Can P pardon the jail term that Smith owe's him? I don't know, that seems mistaken.

Many think (1) is false, e.g., Woltersdorff, I think. See his paper in the Dole & Chignell volume on justice and mercy. It's very hard to embrace both (1) and (2) without making it impossible to solve the paradox of mercy.

Mike's point is also a deep concern for the view; not sure which is the deeper problem.

If this theory is under the supposition that God is real and created the universe, then there is no problem with 1 - 4. God created the universe, including right and wrong. I know that it is common to think about whether God wills something because it is right or if it is right because God wills it, but those two choices seem silly to me. If morality is real in an objective sense, then God created it. And if God exists, created morality, chooses to always be moral, and communicated something roughly equivalent to 1 - 4, then 1 - 4 are fine regardless of any objections anyone might think up.

I don't know what the paradox of mercy is, but I'd like to know (is it that God can't be all-just and all-merciful?).

Anyway, some other worries for the penal substitution theory, as you've sketched it above.

First, there is some tension between (2) ("God, as a just God, must therefore satisfy the demands of justice") and accepting a divine command theory. The tension, obviously, is that if moral obligations are constituted by God's commands, then it is hard to see why justice (which presumably is also in some way constituted by God's commands) is something that constrains God. Now, perhaps one can resolve this by holding that, although moral obligations, etc. are established by God's commands, it is nonetheless the case that, once established, what God has decreed to be a moral obligation is unalterable, and applies even to him. Of course, if you're not a divine command theorist, then this worry doesn't apply to you; it might even give a reason to favor the penal substitution theory--because it makes divine command theory untenable.

The other problem is perhaps with (3) ("because of God's great love for us, God pours his punishment and wrath not on us, but on Jesus, God's only Son"); doesn't God love Jesus/Himself as well? If so, why does His love for us render it impermissible for Him to punish us but not impermissible for Him to punish Jesus, Whom He also loves?

I think Mike's point causes me the most trouble, though. It also seems to be what atheists fasten on when they criticize the doctrine of the Atonement.

on the paradox of mercy: think of ancient views on the requirements of justice. you kill someone, we kill you; you put out my eye, I put out yours. Now comes amelioration based on excusing circumstances: I was trying to save your life and it required putting out your eye. So, if mercy is just an amelioration of excessively coarse-grained requirements of justice, then there is a place for mercy in our moral theory. But then the demands of justice are pretty objectionable. So suppose the demands of justice aren't pretty objectionable. Then there is no place for mercy, since all grounds for it have already been taken into account. So either the account of justice is objectionable and there is a place for mercy, or it isn't and there isn't. See the Woltersdorff argument mentioned earlier for an exploration of the issues here.

Here's an argument in favor of T, or perhaps it is just a near relation to T. Andrew can decide whether it is close enough for his purposes.

Consider a system of vicarious punishment, whereby instead of X being made to suffer as punishment for X's crime, some person related in some way to X is made to suffer. The idea is not that instead of X's being punished, person-related-to-X is punished; the idea is that X is punished not by suffering Xself but by this person related to X being made to suffer. One can imagine pretty much every aim of punishment being met on such a view. Deterrence, retribution (I'd rather suffer for my deed than, say, have my spouse suffer for it, etc.), rehabilitation (perhaps one would be more moved to repentance by others' having to suffer in one's stead, etc.)...

Now ask: What's wrong with such a system? Well, of course the central objection is the injustice of someone else having to suffer for one's own deed. But if the person who does have to do the suffering accepts it willingly, then there is no injustice.

So perhaps the penal substitution view is better formulated not in terms of Jesus being punished in our stead, but Jesus undergoing the suffering in our stead. We are condemned adequately — indeed, one might say more adequately — through Jesus's suffering.

(T) Jesus can satisfy the demands of justice by suffering the punishment for our (humanity's) sins.

There are at least two objections to (T) that I can see. Suppose you badly insult Jones. Sometime after you send Smith to apologize for you. Two cases:

1. You are not even apologetic, but you believe that, nonetheless, your debt to Jones will be repaid if Jones gets an apology from someone.

2. You are apologetic, but you refuse to face the humiliation of having to apologize to anyone yourself.

(1) does not seem to rectify the injustice. So one might object to (T) by claiming that Jesus can defray our debts ONLY IF we actually regret what we did. (2) does not rectify the injustice either, since we are unwilling to incur any penalty ourselves for what we did. So one might object to (T) by claiming that Jesus can defray our debts ONLY IF we are ourselves disposed to pay them if Jesus does not.

A-

I don't know what you mean by 'demands of justice'. It might require people getting what they deserve. In that case (T) is false. The demand of justice cannot be met be a sacrifice made by Jesus since it is unjust both that he is punished undeservedly and that those who deserve punishment are not punished. The key idea is desert. If you run the mile under 4min you deserve praise, but nobody can deserve that praise on your behalf...etc..i think the same goes with every case of desert of F.

Here are some worries.

Claim 1 is false as human children have not sinned, nor have zygotes. Perhaps you mean 'human persons' have sinned. I think this will still be false as people die before committing a sin, just after acquiring personhood, e.g. kids in auto accidents.

Claim 2 will be denied by some divine command theorists, but that is not a problem.

Claim 3 & 4 are terribly odd. Presumably god does not 'pour wrath' but allows something bad to happen, namely that other people harm Jesus. Now, suppose you have a right to punish your kids for scribbling on the wall again, does that mean your neighbor is so positioned? No, they are your kids, not his. Thus, punishment when adminstered must be administered by one with a right to do so. The Fs (i forget) who killed Jesus had no such right, so the demands were not met.

Second, it seems to me that if justice requires X units of badness, then ought to be distributed so that lots of people get some of pile. It's debatable, but not implausible, that distribution of pnishment is a constraint on its fairness. So the demands of justice, whatever that means, requires some equitable distribution across persons.

The key idea is desert. If you run the mile under 4min you deserve praise, but nobody can deserve that praise on your behalf...etc..

No they can't deserve it on your behalf, but they can accept it on your behalf. People accept awards on behalf of others all the time. I guess that's what occurs in the punishment case: someone accepts or receives the punishment on your behalf, but they don't deserve it on your behalf. Still, as you note, there are lots of other problems with this idea.

Re: the paradox of mercy:

Suppose you're a retributivist, but you're not a classical retributivist (a classical retributivist thinks: the state (or God) is morally obligated to punish the guilty as much as they deserve); instead, you're a moderate retributivist like Michael Moore (the legal theorist, not the propagandist) (the moderate retributivist thinks: the state (or God) is morally permitted to punish the guilty as much as they deserve).

In this case, we could rewrite (1) and (2) to read:

(1*) Human beings have sinned, and so justice permits that we be punished for our sin.
(2*) God, as a just God, may therefore satisfy the demands of justice.

Now, if God punishes people to the maximum degree, then he is just. On the other hand, if he doesn't punish people to the maximal degree (which the demands of justice do not oblige him to do), then he is merciful.

Is there a problem with this reformulation? Also, if we accept moderate retributivism, do we even need to reformulate (1) and (2)?

Now, if God punishes people to the maximum degree, then he is just. On the other hand, if he doesn't punish people to the maximal degree (which the demands of justice do not oblige him to do), then he is merciful

Robert,

That does sound intuitive. On the other hand, it can't be that God does not punish the deserving out of mercy but also exacts punishment from a substitute. But this might be beside the point you're making.

Well, if you think that God is permitted but not required to punish to the maximum deserved, then the penal substitution view has no explanation for endorsing (3). At the heart of this view of the atonement is that someone has to pay the price for sin. So replacing (1) and (2) with the more plausible (1*) and (2*) undercuts the needed grounds for (3). That's Woltersdorff's position in the article I cited earlier.

Thought I would mention other problems...

Justice requires proportionality, the idea that the punishment should fit the crime. Illustration: It's wrong to hang somebody for running a red light.

Thus, the allowing of Christ's execution, if it is to meet this proportionality requirement, must be proportional to the sins for which it is a punishment. But it seems too weak. The sins of the many, which include not a few murders, rapes, and other nasty things, deserves far more than a single allowance of an execution. The atonement is thus insufficient.

Similarly, allowing people to burn eternally in hell is disproportionate to any sin they might committ. Thus, allowing people to be in hell is unjust. And for what it's worth, I don't think free rejection of grace will get God off the hook. He could take the bad people to Heaven, show them what they are missing, and nobody would freely spend an afternoon burning...

Finally, the idea of accepting punishment on another's behalf is insufficient for justice. If the idea is that some punishment is deserved and Christ could accept it, and thus, God's allowing Christ to accept it, is sufficient to meet the demands of justice, then we violate a constraint of directionality on justice. The only person who can rightly accept punishment for some state of affairs is the one responsible, morally, for its existence. Even if I accept someone else's debt, which I can, what generates my obligation to repay it is my promise, and punishment would be deserved for violating that promise. But promise breaking would be the only act for which those whose neglect an accepted debt could be faulted. And presumably, allowing Christ to be killed, even if it may be required so that Christ could keep a promise, God would have an obligation to save Christ so that he violated it. Illustration: I ought to violate a promise to repay you five bucks if it means saving someone recently hit by a car...

More thoughts...

Here are some thoughts! I hope they're not un-welcome (and I apologize if they seem rushed):

"So one might object to (T) by claiming that Jesus can defray our debts ONLY IF we are ourselves disposed to pay them if Jesus does not."

You can't forget though that in order accept Christ's propitiation, one must repent, which is essentially an acknowledgment of, and taking responsibility for, all sin one has committed.

People are viewing the viability of 1-4 based on a very human economy of justice. Biblically though, there is much parallel made between monetary debt and the debt of sin. So, we can suppose:

"...you have a right to punish your kids for scribbling on the wall again, does that mean your neighbor is so positioned? No, they are your kids, not his."

But this view is not entirely consistent with what is presented biblically.

If someone, say, owes you $10,000, what you want is repayment. The debt must be reconciled. You would not typically concern yourself with how the debt is repaid. If the person who owes you needs to take out a loan elsewhere, has to sell their possessions, get a second job, or has someone else pay on their behalf, is irrelevant to you. All you are concerned with, is repayment.

In this case, all that is needed for justice, is payment of the debt. Whether or not you personally feel justified, having been wronged, is not the determining factor in whether justice has been dealt or not. Sin is an account with God alone, and since, justice indeed a value for him, we know that he will set things to right, on behalf of the wronged.

Finally, the idea of accepting punishment on another's behalf is insufficient for justice. If the idea is that some punishment is deserved and Christ could accept it, and thus, God's allowing Christ to accept it, is sufficient to meet the demands of justice, then we violate a constraint of directionality on justice.

Again, this is true if sin really is between parties -- which is certainly the human point of view. "You hurt me." "You stole from me."

But if we take the view that sin, even committed against a fellow human, is rather a sin against God (and against Him alone) this problem disappears. I do not think this view is just fancy footwork and in fact we find as far as the Scriptures are concerned, this is very likely the case. Portions of Scripture regarding law and justification between parties can be viewed as a working out of God's desire to bring justice to the abused -- but re-compensation for sin is exclusively expressed towards God. We can view all apology, repayment, and setting-to-right as being ultimately expressed to God.

Jesus can pay debt on my behalf, and the act is not-invalidated. It is unto God alone that I am indebted, and it is He who has established the criteria of repayment.

I have sympathies with many of the points raised.

(a) It is not true that, if a sinner goes unpunished, injustice has been allowed to flourish (where, presumptively, that’s a bad thing). In ordinary human life, sinners get forgiven all the time, and people are reconciled. In fact, unwillingness to forgive and forget, without punishment, is usually taken to be a character flaw in the offended person. So if humans can forgive and forget, why can’t God? Thus I disagree with (1) and (2) in the original formulation; this is also Jon K’s and Wolterstorff’s position, I take it.

(b) I also agree with Mike A’s original objection, which is that punishing B for A’s sins is not just, no matter what B’s status or willingness, etc. Moreover, any punisher who went in for this would be deeply unjust. The problem here is not just that A does deserve punishment and doesn’t get it, but that B doesn’t deserve punishment and does get it.

Two thoughts I’ve had that might make for some progress:

First, the “sin” and “punishment” at issue might be seen as collective, rather than individual. There is considerable biblical support for this. I.e. the punishable sin is attributed to humanity as a whole, and the punishment is dished out onto humanity as a whole, in the legal sense. This is consistent with different individuals within the collective committing the actual infraction and absorbing the actual punishment. Think of the legal situations involving, say, corporations and their employees or stockholders.

Second, part of the problem may involve projecting a retributive notion of justice onto a tradition that has a more restorative view of justice. “Justice” in the Bible (not unlike Plato) has to do with a well-ordered community, not simply with an eye for an eye. In general, as per my first comment, the idea that “justice demands” this or that strikes me as fetishistic. I don’t know how to pursue this further, though.

Andrew,

Here are a couple of additional troubles that seem to have gone unnoticed (perhaps because they do not merit consideration).

1) How can any activity be properly called a "punishment" when it is freely chosen by the individual undergoing it? Suppose I steal a car and a judge sentences me to a punishment of no less than three years imprisonment. It seems reasonable to me that the punitive aspect of this sentence stems entirely from the obligation it presents on my part, i.e., it is a punishment precisely because I cannot help but accept the sentence, it is forced on me.

Now suppose that Mike Almeida steals a car, and the judge gives him a choice, either he can sentence him to three years of imprisonment or he can sentence a willing surrogate to the same three years. Now, being the nice guy I am, suppose that I freely chose to serve as Mike's penal surrogate (perhaps because I think that Mike's three years of imprisonment would be a greater disservice to the philosophical community), in this case have I been punished by the judge? It seems reasonable to me that I have not been punished at all, I simply chosen to serve my friend (or my community) by the constraining of my freedom.

If this is a more or less accurate picture, then it seems to me like "punishment" requires an imposition of agent-neutral demands that cannot be satisfied in the case of the penal surrogate.

2) There is the problem loosely captured by Stump and Adam's arguments for individualist constraints on theodicy. A thumbnail of the point goes like this: in order for God to be justified in causing an indivudual to suffer, it must be necessarily the case that there is some good obtained by the suffering of the individual which is necessary for the the flourishing of the individual and which the individual cannot get my another means at all. In this case, in order for God to be justified in causing Jesus to suffer in on our behalf, there ust be some good that Jesus gets from the suffering, which cannot be gotten in any other way at all. It's not clear what this good might be in the penal surrogate case.

Christian, you write,

Justice requires proportionality, the idea that the punishment should fit the crime. Illustration: It's wrong to hang somebody for running a red light.

There is something right about the proportionality requirement, but it is very hard to say what. Ted Sider says somewhere in his nice paper on hell that proportionality requires that any two people S and S' that are nearly the same in morally relevant respects must be treated in nearly the same way. Is that true? If S wins wins the Kentucky Derby by .0001 of a second, and S' comes in second, should they be treated in nearly the same way? The fact is that they certainly are not treated the same way, despite being nearly the same, and no one sees any injustice in it. If X is irredeemably evil and X' is borderline irredeemably evil, it seems to me they can be treated in very different ways. It's true that X cannot be redeemed, it is not true that X' cannot be redeemed, despite the fact that they are in nearly the same moral state. So I'm not sure what the requirement of proportionality comes to, once it gets spelled out.

Andrew, you say,

Another way out is to argue that justice does not demand apologies.

I don't think I said it did. I think I said that being apologetic (we might better say, genuinely regretting what one did) is necessary for justice if someone else is being punished in your place. I don't think it is necessary if you are the one being punished.

"How can any activity be properly called a "punishment" when it is freely chosen by the individual undergoing it?"

This again reflects one view of what justice might be, and might not consider it thoroughly enough from the Divine point of view. I might be wrong, but it is a different economy than we are typically used to working in.

Consider that in God's being just, He demands that a debt (sin) be repaid. But on what basis can assert that the sin is inextricable from the sinner? When did the two become an indivisible entity, and why is this view justifiable? As humans we view punishment as necessarily having to be poured out on the offender because we feel that both are tied inseparably -- but the Scriptures suggest that perhaps the offender is not his offense. We want vindication! But how does this translate to the Divine court? I think there's plenty of room to hold the view that sin need not necessarily be a "possession" of the sinner. That is, it can be transfered (again, much like debt). We see this theme elsewhere in the Scriptures, for example, in the Judaic tradition of passing sins onto an animal for sacrifice, or the concept of generational sin.

I'm just saying, the Biblical opinion of justice does not adhere strictly to what our Western court-room conception of justice is. If we're going to be consistent in evaluating this view, we need to consider the legacy of Scripture, and not Christ in isolation.

"in order for God to be justified in causing an indivudual to suffer, it must be necessarily the case that there is some good obtained by the suffering of the individual which is necessary for the the flourishing of the individual and which the individual cannot get my another means at all."

Saying God needs to be justified I think is to force Him into the kinds of constraints that we are used to applying to each other, namely, that His actions need justification (and in accordance with anything that we can fully understand). If He's God, He is His own justification. But for the time being, granting that this might be the case, I'm not sure why it follows that anything need be a justification other than God's desire that sin to be accounted for.

Maybe I'm missing something, but where is the connection that Jesus must somehow get something out of it as the surrogate?

Now, if it is indeed a valid criteria, consider this: Christ was doing the Father's will. If Christ had decided, "Nah. I don't wanna do that," He'd have been disobedient, thus sinful and not God. Further, He would not a valid surrogate, as blamelessness is/was a requirement for this kind of undertaking.

So if anything could be gained on the part of Christ from the exercise (again, apart from God's demand for justice being satisfied) it was in accordance with the will of the Father, and therefore continued harmony and fidelity within the person of God. Anything less than Christ not fulfilling the Father's will for the atonement of man and accounting of sin would have equated to God not being God.

"It may be that apology (and confession and a contrite heart and repentance) are necessary for reconciliation in a relationship, and that is different from satisfying the demands of justice. What do you think?"

FWIW, I think this is a great point. I tried to allude to something similar in my first comment. Out concept of justice is so caught up with our desire to be validated, and as I said above vindicated. We'd do well to try and separate out what the human view of "justice" is, and how this might differ from what God's expressed opinion might be (that is, what the Scriptures say). We might not agree with what the Scriptures say, but if we're evaluating the viability of the penal substitution of Christ, it needs to be in the context of what the Scriptures say about justification elsewhere -- we need to try and work within its economy.

Sorry if I'm stating the obvious!

"I said that being apologetic (we might better say, genuinely regretting what one did) is necessary for justice if someone else is being punished in your place."

Again, to have Christ's sacrifice credited to your account, this is a requirement in the form of repentance.

I'm going to admit up front that I haven't read much of what came above--though I did scan it and search for some key terms using Ctrl+F), so if I repeat any argument somebody just censure me.

There are two big problems with the the Penal Substitution view. First note that Andrew doesn't say what he thinks it's a theory *of*. It is typically taken to be a theory of the atonement (and it's no urban legend that Wyclif made that term up "onement" to translate "reconciliato" then I think Tindale extended it to "atonement"). At the broadest level of generality we are reconciled to God by repenting of our sins and being forgiven. This is the first problem with the Penal Substitutionary Theory of the Atonement: there is no forgiveness anywhere in the picture. To forgive is to *cancel* the debt, not to have it paid by some alternative route. Salvation comes from declaring spiritual bankruptcy, not by having a winning lottery ticket. (For more on this see Eleonore Stump's excellent "Atonement According to Aquinas" in Morris, ed. _Philosophy and the Christian Faith_, UNDP, 1988. See also Swinburne's excellent stuff in _Responsibility and Atonement_, also Robin Collins (of FTA fame) has (or did have) some good stuff on his website).

The second problem is that it seems to entail the disjunction of limited atonement or universalism both of which are false. (Hope I'm wring about the last one but I doubt it). Either Jesus died for all or not-all. If for all then either all are saved or not all are saved. If all are saved then Universalism. If not all are saved then Christ died in vain for some. (Those who like PS are very unlikely to accept this.) But if for not all then Limited Atonement.

Oh I meant also to recommend Swinburne's "The Christian Scheme of Salvation" in the same Morris volume as the Stump article.

I think people sometimes get confused into thinking they believe the Penal Substitution theory by the fact that they like the story of Damon and Pythias (thanks to my 9-year-old daughter Fiona for the reference, I had thought it was Castor and Pollux (yeah, she's home-schooled)).

That's the story (most famously told by Cicero) where Pythias had defied the totalitarianism of Dionysius King of Syracuse (sounds funny to me know living in upstate NY). He was to be imprisoned in a cave but asked to be allowed to return home to say his goodbyes. The King agreed to the deal whereby if Pythias didn't return his best friend the willing Damon would be killed. Pythias made it back just in time (after overcoming many trials). Needless to say the king was moved by this token of affection and pardoned both parties.

It's pretty clear that this story favors some theory other than the Penal Substitution theory. Some will protest that it implies that God needs to be moved to compassion but that is a non sequitur. You might also read Philip Quinn's article on Abelard's theory of atonement in _Reasoned Faith_ edited by Eleonore Stump (Cornell Press, 1993).

Other good reading is David Brown's entry in the Cambridge Companion to Anselm, Quinn also has an article on Aquinas on Atonement followed by "Atonement and Justification" by Stump in the Feenstra and (Cornelius) Plantinga volume _Trinity, Incarnation, and Atonement_ (UNDP 1989), finally Quinn has a reply to Swinburne in the volume of essays in the latter's honor.

Mike,

I don't believe that people deserve punishment. You're right that cashing the proportionality claim out is hard. But I think the Jesus case is like the red light case. Assuming justice requires proportionality we should think Christ's death couldn't atone for people's bad deeds. We don't need to take a stand on a theory of proportionality.

But how about this for proportionality: Welfare comes in units of welfare. A punishment is proportional just in case the units of disvalue to the one punished = the units of value the one punished deprived from others. This is eye for an eye.

Oh, and I'm sketchy about the K. Derby bit. In a sensitive state of mind I think the winner and second place finisher deserve similar praise, but that we are knuckleheads to treat them differently qua what matters to their similarities.

I like Taylor's point above. If x freely accepts undergoing some bad thing without being coerced, could that thing undergone be a punishment for x? You suggest No. Here's a rejoinder: I can freely undergo that bad thing for my sake or for yours. If it is for my sake, then it is not punishment. If it is for yours, it can be a punishment. I can freely accept something I wouldn't prefer to have. Thus, supposing Christ endured the suffering for the sake of others, doing so could still count as punishment.

Christian, your worries about the atonement seem to be here,

The sins of the many, which include not a few murders, rapes, and other nasty things, deserves far more than a single allowance of an execution.

I'm not a theologian, but I don't think that the execution was the payment. There was presumably unfathomable suffering in his taking on our sins. Please don't ask me for the metaphysics of taking on sin ... I don't have a clue.

A punishment is proportional just in case the units of disvalue to the one punished = the units of value the one punished deprived from others.

This can't be right. You can have two agents S and S' whose moral states are not definitely different (so one did not do much worse than the other) but where S' should be treated much worse than S.

I like Taylor's point above. If x freely accepts undergoing some bad thing without being coerced, could that thing undergone be a punishment for x?

I'm not sure where I said it couldn't. In any case, it is not a punishment for Christ in Christ's case. It is a punishment for us, but he is taking it on.

Mike,

"I'm not sure where I said it couldn't." No. I meant that for Taylor. Not you. The way I indented and whatnot made it seem otherwise. Oops.

"Please don't ask me for the metaphysics of taking on sin." So, what's your metaphysics of sin:)

If it were reasonable to say Christ's death was so bad that it atoned for everything bad ever done, then, well, I live in cloud cuckoo land. So much the worse for the metaphysics that implies it.

"This can't be right." Eh? Maybe you're right, you're good at the counterexamples, but do you have one in mind?

"This can't be right." Eh? Maybe you're right, you're good at the counterexamples, but do you have one in mind?

No, there are counterexamples. It can be that S is not definitely better than S', and yet S' is definitely irredeemanbly evil and S is indefinitely irredeemably evil. In that case S' is an irredeemably evil person and S is borderline irredeemably evil. That can be true event if S and S' do not differ by much more than a few uttered obscenities. God can save one, but cannot save the other; one is beyond redemption and the other not. So a few obscenities can make all the difference.

I have a quick and probably unlearned question.

It seems to me that the New Testament writers (and much of theology) emphasize the fact that there just is an imbalance between human sin and God's choice to act as if Jesus' death atones for it. This also seems to be the essence of forgiveness - God acting *as if* the sin had not occurred - when objectively it has. There is no objective "justice" involved. And it seems to me that this is the only kind of fact that could ground meaningful thankfulness to God beyond the thankfulness involved in normal etiquette (i.e. we aren't thankful for the obligatory actions of others in as meaningful a way that we are for supererogatory actions).

A related note: Wouldn't many of the above objections to penal substitution be even stronger for the sacrificial system established by God for Israel? It seems like the death of an "unblemished lamb," for example, is even less proportional than Jesus for what is deserved by human wickedness. So if we take that institution as God-ordained, then we probably have to come up with a very ancient-sounding metaphysics of sin (which is notably avoided by some here!). There really seems to be a kind of "getting rid of" sin, like it is some kind of substance.

Christian,

I've also wondered how Christ's death could be sufficient to atone for all the wickedness in the world, etc. You're right: it seems one mere execution shouldn't do the trick here.

But here's a different story that I've heard that might help. Paul says somewhere that "he who knew no sin (Jesus) became sin on our behalf that we might become the righteousness of God in Christ." And when Jesus "became sin" for us on the cross (whatever that amounts to), it actually ruptured (in some sense or other) his relationship with the Father. Hence, you get Jesus on the cross crying out "Father, Father, why have you forsaken me?" -- a question that seems to indicate a broken relationship.

But Jesus had always had a perfect relationship with the Father, being (in some sense or other) eternally begotten of the Father. Indeed, this relationship was one of perfect love and harmony, etc. So when Jesus "became sin" for us (or, as Mike's said, "took on" our sin), he experienced the full alienation from the object of his love that sin brings about. Said a bit differently, it wasn't merely that he was in physical pain (though he was), in some deeper sense he was - shockingly, unnaturally - alienated from God. And since, for Christians, God is something akin to the Good, he was alienated from everything Good. Which is to say that he was in as bad as shape as it gets!

Anyway, I don't claim to be able to explain much about anything I just said. And maybe this is wrong on various theological grounds. But it makes more sense than the idea that all the work was done by a mere execution.

Mike,

I don't follow this: "That can be true even if S and S' do not differ by much more than a few uttered obscenities."

On that hypothesis I would say one is definitely better than the other. But my intuitions about "borderline" cases are murky. I could be argued into thinking I'm wrong, but it seems that S is better than S' iff S is definitely better than S'. Perhaps we cannot know which is better, but that's another story.

"It seems like the death of an "unblemished lamb," for example, is even less proportional than Jesus for what is deserved by human wickedness. So if we take that institution as God-ordained, then we probably have to come up with a very ancient-sounding metaphysics of sin (which is notably avoided by some here!). There really seems to be a kind of "getting rid of" sin, like it is some kind of substance."

In part, I think you're right. Is justice all a matter of "balancing the scales" with God? Bulls and Lambs would seem to suggest, not really. One thing about mercy, is that it is undeserved, and just as God is just, so too is He merciful. We see its ultimate expression of mercy in propitiation through Christ. But on proportionality....

"I've also wondered how Christ's death could be sufficient to atone for all the wickedness in the world, etc. You're right: it seems one mere execution shouldn't do the trick here."

Oh yes. It very much should. The term "mere" reveal, methinks, a particular point of view regarding Christ, namely, His Godhood (or lack thereof).

"...it actually ruptured (in some sense or other) his relationship with the Father. Hence, you get Jesus on the cross crying out "Father, Father, why have you forsaken me?" -- a question that seems to indicate a broken relationship."

Consider, for the moment, that Christ is not just "some guy," but rather, that He is God Himself.

His "penal substitution" is not at all like my going to jail for the theft or murder a fellow human committed, in terms of equity.

I should think that, if Christ is in fact fully God (though fully man) then in fact His sacrifice is infinitely greater in value than the sum total of sin accumulated by all of human kind.

That is, not only would the debt be paid in full, but there would be an abundance of currency left over. Is there anything Biblically that would account for this? I believe yes. We are told of inheritance -- that those who choose Christ will be co-heirs with Christ. Co-heirs of what? Hmm...

But yes. There was perhaps no more existentially problematic and devastating event than the severing of the Godhead for a moment in time. There could be no greater sacrifice than this. Yes, the deficit would have been amply accounted for by such a sacrifice. Accounted for, and then some.

Christian,

Take any vague predicate, 'x is red' for instance. Suppose you have continuum many objects (or just space-time points on a spectrum) that instantiate every shade of red. There will be on the spectrum two adjacent points P and P' such that (i) P' is not defnitely more red than P, (ii) P' is red and not on any borderline of red, and (iii) P is on some borderline of red. The same goes for vague moral predicates. There will be a single obscenity that distinguishes those that are irredeemably evil (IE) and not on any borderline of IE, and those that are on some borderline of IE. One is not definitely more evil than the other, yet one is [IE and on no borderline of IE] and the other is [borderline IE].

A couple of remarks.

I. Claim (1) should be weakened. Justice does not demand appropriate punishment. But appropriate punishment is fitting. This medieval term, "fitting", is the right theological term for these discussions. God has a reason to punish, but he may also have reasons not to punish. Even weakened in this way, claim (1) supports the penal substitution theory. Indeed, in this way it should heighten our adoration of God--we can thank God not only for having saved us, but for having done so in such a fitting manner, despite the terrible cost.

Here is an argument that justice does not demand punishment. Consider the flip side: reward. If someone has done you a good turn, that deserves a good return. But while a failure to return a freely offered favor is less than fitting, it is not wrong or contrary to justice. If it were wrong, then by doing good to someone, we would be binding her with the ties of obligation, and then we would have the nightmarish society of the Gands with their obs (yes, Frank Russell the author thought it utopian, but he's wrong) where genuine altruism is night impossible and attempts at it are rebuffed.

II. Christian's point that Christ didn't suffer enough to substitute for all the sins of the world forgets that just as the greater joys in life are spiritual so the greater sufferings may be spiritual. The virtuous person's feelings in good observing conditions tend to reflect the reality of the situation, except insofar as the feelings are repressed by special circumstances. On the side of the exception, we have the dreadful feeling of abandonment by God. On the side of the reflection of the reality of the situation, Christ understood the fact of evil better than anyone, was in particularly "good observing conditions" for evil on the Cross, stood in solidarity with the victims of evil, and loved every sinner and suffered at the way sinners ravaged their very own souls. Just as the virtuous person rejoices in the virtue of another, so too she tends to be pained at the other's vice. But barely understand the enormity of the least sin as an offence against God. But Christ understood it, and we may hypothesize that on the Cross that understanding would cause proportionate suffering. After all, think of how much a sensitive person suffers whe she has inadvertently seriously hurt someone that she loves. Christ would, on this hypothesis, feel the even greater pain that someone who has deliberately seriously hurt those that she loves ought to (but often do not) feel. He stood with sinners and victims, in compassion for both. Beyond that he suffered at those who would reject--whether permanently or temporarily--the grace he was bringing, the grace of forgiveness. If one wants, one can add (the Tradition generally asserts this) that even as a man he had knowledge of all human affairs, past, present and future, which would heighten the suffering. (Granted, it could also bring joys. But joys need not lessen suffering. And perhaps he wasn't in "good observing conditions" for goods.)

III. I think Trent's point about cancelation making forgiveness impossible is a good conceptual point, but it breaks down when the cause of the cancelation and the forgiver are the same person. But it is the same Christ whose sufferings cancel our debt and who thereby forgives our debt. (One might say that the Father forgives as well, and there might be some Trinitarian issues that endanger what I say, but I think what I say is basically right.)

Mike,

I don't claim to know much about the correct characterization of predicates the satisfaction of which admits of borderline cases. From the hip I'm skeptical of the 'definitely" and 'indefinitely' modifiers. They seem to me to generate the same problems they are appealed to to solve. If we doubt there is a sharp cutoff between red and non-red, then we should doubt that there is a sharp cutoff between definitely red and indefinitely red. Then I see no reason to introduce the modifiers.

Anyway, if (i) P' is not definitely more red than P, then that is because either (ia) P' is definitely less red than P, or (ia) P' is the same colorof red as P. If (ib), then ((ii) and (iii) are jointly incompatible. If (ia), then P is not red, and given that P and P' are adjacent, (ii) is false. Either way, the description is jointly incompatible with the assumptions.

Given proportionality, If S is causes n amount of suffering and S' some amount distinct from n, then S deserves n amount of suffering and S' that amount distinct from n. Nothing about borderline cases casts doubt on this. You seem think that S and S' might not be definitely different wrt the suffering they cause, but that one can deserve a punishment of a much greater amount than the other. I just don't see this. They must cause different amounts of suffering to deserve different amounts of punishment. If there is a threshold, then that's a different story. I deny that desert has thresholds. It comes in degrees. And I deny that God can meet the demands of justice by selecting an arbitrary threshold and then giving one who barely does not meet it some punishment disproportionate to the one who barely does. That's paradigmatically unjust.

Alex and Patrick,

I appreciate your strategies. Describe some feature that Christ's suffering on the cross involves and say that it, in addition to the physical pain, is proportionate to the sins of all persons distinct from Christ. If there was such a feature, I agree, proportionality would be met. Patrick, you point to a "ruptured relationship with God" as one such feature. You mentioned "being alienated from God and all that is Good" as another one. But are such features sufficent? I say no. Christ's alienation and ruptured relationship was temporal and He, by hypothesis, spends eternity in Heaven with God, thus being being in a non-ruptured and non-alienated relationship. Moreover, part of people's suffering involves the very same thing, and there are many more of them, i.e. there are many people who alienated and rupture their relationship with God and that is a sin. Finally, these are simply cognitive/emotional bad features, but no amount of cognitive distress adds up to the cognitive distress caused by others.

Alex points to the feature of being in "solidarity with the victims of evil" and suggests that, maybe, "understanding would cause proportionate suffering". I would deny that understanding or solidarity with sinners is sufficent. Understanding that someone suffers, if one is sensitive enough, causes some suffering. But many sensitive people, in solidarity with others, are victims in this. And I would suffer "much" less by virtue of watching in sensitive horror someone getting tortured to death than that person who is being tortured to death. Moreover, Christ lacks what those victims have, understanding from the their own, first person, point of view. Understanding that there is suffering endured by x, y, and z, is not the same as x's understanding that x has suffered through no fault of his own, and y's understanding that y has suferred...summing up the interpersonal cognitive sufferings, of each victim, far outweighs the cognitive suffering of one, even if He could appreciate the "grand scheme" of things.

If we doubt there is a sharp cutoff between red and non-red, then we should doubt that there is a sharp cutoff between definitely red and indefinitely red.

There is no sharp cutoff point, and we don't want one. This is why we have higher-orders of vagueness.

if (i) P' is not definitely more red than P, then that is because either (ia) P' is definitely less red than P, or (ia) P' is the same colorof red as P.. .

False. P' might be indefinitely more red than P.

You seem think that S and S' might not be definitely different wrt the suffering they cause . ..

Actually, I know it. I can attach you to a device that increases pain in increments sufficiently small that your mental state at M is not definitely worse than your mental state at the adjacent state M+1. Since you are not omniperceptive you are not sufficiently sensitive to perceive a definite difference at each increment. You will perceive definite differences at larger increments only. So certainly there might be no definite difference between two torturers T and T' in the amount of suffering they impose, since T might impose M and T' might impose M+1.

Mike,

"There is no sharp cutoff point, and we don't want one."

Then I don't know what you mean by 'borderline' in (ii) and (iii).

"P' might be indefinitely more red than P."

I also don't know what you mean by 'indefinitely more red than'. These are technical terms and I'm ignorant of the literature in wchich they are used.

"So certainly there might be no definite difference between two torturers T and T' in the amount of suffering they impose, since T might impose M and T' might impose M+1."

I'm writing a paper on this case. So I need to think about this point you're pushing. I agree M is indistinguishable from M + 1 for a possible subject. So agreeing I feel no pull to accept that 'indistinguishability' implies 'no definite difference'. I may not be able to distinguish between M and M + 1, but one is closer to M - 1 and the other to M + 2, which makes for a definite difference, even if I cannot tell, from the inside, that difference obtains.

Do you think any of this matters to the principle point, that proportionality* (proportionality suitably modified in light of your worries) is violated with the alleged atonement?

Mike,

"There is no sharp cutoff point, and we don't want one."

Then I don't know what you mean by 'borderline' in (ii) and (iii).

"P' might be indefinitely more red than P."

I also don't know what you mean by 'indefinitely'. This is a technical term and I'm ignorant of the literature in which it is stipulated.

"So certainly there might be no definite difference between two torturers T and T' in the amount of suffering they impose, since T might impose M and T' might impose M+1."

I'm writing a paper on this case. So I need to think about this point you're pushing. I agree M is indistinguishable from M + 1 for a possible subject. So agreeing I feel no pull to accept that 'indistinguishability' implies 'no definite difference'. I may not be able to distinguish between M and M + 1, but one is closer to M - 1 and the other to M + 2, which makes for a definite difference, even if I cannot tell, from the inside, that difference obtains.

Do you think any of this matters to the principle point, that proportionality* (proportionality suitably modified in light of your worries) is violated with the alleged atonement?

I may not be able to distinguish between M and M + 1, but one is closer to M - 1 and the other to M + 2, which makes for a definite difference, even if I cannot tell, from the inside, that difference obtains.

I don't think there can be any defnite differences in my well-being or my suffering if "I cannot tell" the difference. I'd otherwise be committed to imperceptible decrements in well-being, and there aren't any such decrements (as far as I'm concerned).

Do you think any of this matters to the principle point, that proportionality* (proportionality suitably modified in light of your worries) is violated with the alleged atonement?

It is central to counterexamples to various prniciples of proportionality. So it is relevant to formulating your worries. But, yes, it is somewhat beside the general thrust of the thread.

"I don't think there can be any definite differences in my well-being or my suffering if "I cannot tell" the difference."

That's a strong thesis. Would your well-being be same if you were in the Matrix? And would you deny the existence of posthumous harms and benefits? Some would count those as costs. Anyway, I'll give this more thought.

"I don't think there can be any definite differences in my well-being or my suffering if "I cannot tell" the difference." That's a strong thesis. Would your well-being be same if you were in the Matrix? And would you deny the existence of posthumous harms and benefits? Some would count those as costs.

We were talking about increments in pain. If my mental state is M and you increase an electric charge by one (small) increment and my mental state is then M+1, then if I cannot tell any difference (not even an indeterminate difference) between M and M+1, then you have no decreased my well-being or caused me to suffer any more by that single increment. I don't regard that as a strong claim or position. The alternative requires the recognition of imperceptible decrements in well-being. None of this involves denying the relevance of future percpetible pains, say, from smoking excessively or from irradiating someone without them perceiving it now.

I was simply generalizing. You said that if one cannot tell the difference between M and M + 1, when one is at M, then increasing M to M + 1 cannot make a definite difference to one's welfare. The generalization was that if one cannot tell the difference between M and N, when one is in M, then changing M to N cannot make a definite difference to one's welfare.

The logical form is different. You needn't accept my generalization. Though the generalization may be thought to underly one's intuition in the pain case and, as such, would be questioned by Matrix scenarios.

Here's a "solution" to Quinn's case: Welfare at t is intrinsic to S. Consider S in some arbitrary M (that is bad). S can say "this sucks". He thereby demonstrates what M instantiates. Later, a few days later, S finds himself in M + 1, a different state, and demonstrates what it instantiates by saying "this sucks". In each case S demonstrates a different property, though, if S were to move from one to the other he couldn't tell the difference. Given that they are intrinsic, the welfare each contributes to S needn't effect one another. And given that the the states are different, and even though S couldn't tell the difference between them, one is worse for S than the other. It is 1 units worse.

And given that the the states are different, and even though S couldn't tell the difference between them, one is worse for S than the other.

This assumes what I think is untenable. In Quinn's case we have increments (or decrements) in pain. It cannot happen that there is no perceptible difference in one's pain-state between state kn and kn+1 AND the agent in kn+1 is worse off than he was in kn. He's not worse off, since you cannot be worse off in the pain you're suffering unless there is a perceptible decrement in pain. That's what makes it difficult to solve. So, although you might make an omniperceptive agent worse off with any decrement in pain (however small) you cannot make a non-omnipercpetive agent worse off in this way. On the other hand, I think Quinn's puzzle is possible, since it does have a model in supervaluation semantics. It is also solvable in that semantics, in my view.

I can't follow the other, generalization, point. I have no idea how skeptical hypotheses are relevant here.

And given that the the states are different, and even though S couldn't tell the difference between them, one is worse for S than the other.

This assumes what I think is untenable. In Quinn's case we have increments (or decrements) in pain. It cannot happen that there is no perceptible difference in one's pain-state between state kn and kn+1 AND the agent in kn+1 is worse off than he was in kn. He's not worse off, since you cannot be worse off in the pain you're suffering unless there is a perceptible decrement in pain. That's what makes it difficult to solve. So, although you might make an omniperceptive agent worse off with any decrement in pain (however small) you cannot make a non-omnipercpetive agent worse off in this way. On the other hand, I think Quinn's puzzle is possible, since it does have a model in supervaluation semantics. It is also solvable in that semantics, in my view.

I can't follow the other, generalization, point. I have no idea how skeptical hypotheses are relevant here.

(More problems with posting this!)

I reckon we are at a stalemate, again. You push that "you cannot be worse off in the pain you're suffering unless there is a perceptible decrement in pain" and I deny it. You say P and me not-P. Your view makes the disvalue of a pain state extrinsic, but I think it is intrinsic. There is an example that supports my view: Consider S and Dup-S (a duplicate of S). S exists and is in M for 1 minute and then goes out of existence. Dup-S exists and is in M + 1 for 1 minute and then goes out of existence. I say S had a better life than Dup-S. You say no. I don't know where to go from here. But again, you raise a nice challenge, and again, I'l think about it.

Christian,

I had to login to read your comment. It didn't post for some reason (or, it has not posted yet, as I'm writing this). There are those who believe that there can be imperceptible decrements in well-being, but they are forced to embrace Parfit's Repugnant Conclusion. If you're given the option of imposing 999 decrements on one person (let that be some severe pain), or 1000 decrements, one each on 1000 persons (let one unit be imperceptible) then you should choose the former. After all in the latter case we have more pain experienced than in the former case. But it's beyond credibility that you should impose severe suffering when you could have imposed fully imperceptible "pain".

Wonderful puzzle! Whereas I would say Parfit's repugnant conclusion is not at all repugnant, but true, in your scenario I do feel the pull that it is "beyond credibility". But here is one for you: You can impose 2 decrements of pain (so that it is just barely perceptible) to one person, or you can impose 1 decrement of imperceptible "pain" to each of one million people. Isn't it beyond credibility to choose the latter?

...you can impose 1 decrement of imperceptible "pain" to each of one million people. Isn't it beyond credibility to choose the latter?

Here's what you're asking me. I can push button A and cause S's mental state to change from 0 pain experienced to slight pain experienced. Or, I can push button B and cause the mental state of a million people to move from 0 pain expereinced to 0 pain experienced. No change in the pain they experience! Of course I choose B!

"Of course I choose B!"

Okay. I don't feel that way. Maybe I'm wrong. But it seems to me that if you move them up from 2 to 3 you make them worse off. For if someone else moves them from 3 to 4, they will feel it. And that just takes a gain of 1 unit rather than 2, and so less of a change than it would otherwise take. And so you make a million people more susceptible to a pain than they would otherwise be. And that seems to me to be bad.

The Parfit argument assumes the utilities across persons are additive, a dubious assumption. (I can make the worry precise. There many ways of combining utilities between persons. I might add the utilities straight as they come. Or I might first raise the utilities to the power p, then add them up, and then raise the sum to the power 1/p. There is an infinite variety of ways of combining utilities.)

I am not even sure utilities within persons are additive in any meaningful way.

There are standard models in economics for summing utilities (taken as measures of preference-satisfaction). But the fact is that you don't need to sum utilities to get the Repugnant Conclusion. You can get it with other social choice principles. There are serious worries with summing utility for hedonistic forms of utilitarianism, but Parfit is already speaking to a crowd that is utilitarian. If you cannot sum utilities, their worries are much larger than the Repugnant Conclusion. You've solved the problem of that Conclusion (out of the pan) but you can't sum utilities (into the fire!).

Alex or Mike,

Any references? That worry is entirely new to me.

Christian,

There is an enormous literature on interpersonal comparisons of utility. Peter Hammond has a paper here http://www.stanford.edu/~hammond/icuSurvey.pdf that is pretty readable. K. Binmore has a few papers on this, too. I don't know if you have it, but Prentice-Hall (Foundations of Philosophy Series) has a nice introduction to Philosophy of Economics where this problem is introduced (Chps. 2 and 3).

Mike:

Which other plausible social choice principles give you the repugnant conclusion?

By the way, it seems important to separate the question of interpersonal comparison, which is plainly sometimes possible (e.g., if the two people are exact duplicates of each other), and the question of additivity.

Thanks Mike. It'll be interesting to see how non-additive views can make sense of the fact that it is worse to kill two people than one, all else equal.

One can have non-additive ways of combining. One can say that two people killed is worse than one without saying that it's twice as bad.

Actually, killing is going to automatically be a bit tricky on a utilitarian account. For instance, it won't be the case that two people killed is twice as bad as one killed, since that will depend on what sorts of futures the two are deprived of.

Which other plausible social choice principles give you the repugnant conclusion?

Lots of them will. Take standard forms of the Pareto Optimality principle. In the case where 1000 get one unit of pain, it is not possible to improve the 999 people without worsening one. So it is P. optimal. Bad news.

On the other score, I was assuming that your worries about additivity derived from worries about interpersonal comparisons of utility. Maybe you have other worries.

Mike,

My worry is that there are many functions that can be used to compute a total utility. The function f(U1,...,Un)=U1+...+Un is only one of these. All these functions preserve domination relations and are invariant under shuffling.

The most obvious example, which works if the utilities are non-negative (otherwise, we just shift them so they are, or just take p to be odd), is f(U1,...,Un)=U1^p+...+Un^p.

You get normatively different results for different combination functions.

You get normatively different results for different combination functions.

You get importantly different cardinal rankings, yes? But linear transformations preserve all of the relevant relations we want among preferences. I'm wondering why you think any transformation is as defensible as any other. Maybe I'm not taking your point.

Mike,

Here was our case, "I can push button A and cause S's mental state to change from 0 pain experienced to slight pain experienced. Or, I can push button B and cause the mental state of a million people to move from 0 pain expereinced to 0 pain experienced. No change in the pain they experience! Of course I choose B!"

Here is a counterexample, off the cuff, suggested to me by Mike Huemer.

Act A...move person s from 0 units to 2 units or act B...move 1 million people from 0 units to 1 unit. You say choose act B.

Next, you are faced with another choice.

Act A'...move person s' from 0 units to 2 units or act B'...move the above 1 million people from 1 unit to 2 units. If you're reason for choosing Act B above is right, then you should choose Act B' now.

The consequence is that, instead of making two people just barely worse off, you make 1 million people barely worse off. This by just making two choices rather than one. It seems to me like a counterexample. The view I'm pushing doesn't have this consequence.

Alex: You said, "One can have non-additive ways of combining. One can say that two people killed is worse than one without saying that it's twice as bad."

Suppose they are twins, duplicates, or people of the same age, build it all in. I say it's bad for a view if it is not twice as bad to kill two than one. Having read a little a bit additivity no, I can't think of any problems for it, not yet anyway.

Here is a counterexample, off the cuff, suggested to me by Mike Huemer.
Act A...move person s from 0 units to 2 units or act B...move 1 million people from 0 units to 1 unit. You say choose act B.
Next, you are faced with another choice.
Act A'...move person s' from 0 units to 2 units or act B'...move the above 1 million people from 1 unit to 2 units. If you're reason for choosing Act B above is right, then you should choose Act B' now.

I'm committed to no such thing. By hypothesis 2 units is noticeable pain. Mike seems to think that if 2 units is not noticeably worse than 1 unit, and 1 unit is not noticeably painful, then 2 units is not noticeably painful. But that's wrong. It can be true that 2 units is not noticeably worse than 1 unit, 2 units is noticeably painful and 1 unit is not.

Maybe this will help.

1. It is undeniable that there is some increment i (in electric current, atmospheric pressure, or whatnot) that is not painful at all.

2. It is equally undeniable that there is some number of increments (n x i) that is noticeably painful.

3. It is therefore undeniable that an agent moves at some point in the sequence from a state ((n-1) x i) that is not noticeably painful to a state that is noticeably painful (n x i) in a single increment i.

I'm saying nothing new, but it helps enormously to have some familiarity with some semantics of vagueness. It is easy to model (1), (2) and (3) in any semantics for vague predicates. I've modeled similar situations in supervaluation semantics, which is what I prefer to work in.

Hi Mike,

By the way, and I know we know this, but my comments should not be taken to reflect Huemer's opinion.

You said "It can be true that 2 units is not noticeably worse than 1 unit, 2 units is noticeably painful and 1 unit is not."

Agreed. The counterexample does not depend upon denying that, it assumes it. The idea is that if you indeed think that a move from i to i + 1 is okay, since it does not make one worse off, then you should think that a move from i + 1 to i + 2 is okay, since it does not make a person worse off. Do you deny this? If so, how could you deny this? We were assuming that the agent cannot tell the difference between adjacent settings. That is why a move up by 1 unit does not, according to you I thought, make the subject worse off. Now it sounds as though you think some moves by one increment can make a person worse off, but some moves by one increment do not.

Before you said "Or, I can push button B and cause the mental state of a million people to move from 0 pain expereinced to 0 pain." Earlier you said, "But it's beyond credibility that you should impose severe suffering when you could have imposed fully imperceptible "pain"."

Why would you think these things if you didn't think that a move up, by one unit, does not, and always does not, make a person worse off? In each act, that condition is satisfied even though *both* acts, considered as one act as it were, does make a person worse off. I just don't get it.

I agree with your (1), (2). I don't understand (3). Perhaps (3) is supposed to capture the idea of a threshold. But that is inconsistent with the setup if we are assuming that for any adjacent settings, moving from one to the other of the adjacent settings is not a move from an unnoticeable pain to a noticeable one. That is stipulated in the case.

You have a semantics of vagueness in mind. I don't think the case involves vagueness at all. I think the case involves either (A) a jointly incompatible description or (B) a case in which one can be made worse off even though one does notice that one is being made worse off.

An impasse, perhaps.

. . . The idea is that if you indeed think that a move from i to i + 1 is okay, since it does not make one worse off, then you should think that a move from i + 1 to i + 2 is okay, since it does not make a person worse off.

Christian, I can't see how you say that (above), and agree, as you do, to this (below),

You said "It can be true that 2 units is not noticeably worse than 1 unit, 2 units is noticeably painful and 1 unit is not."

After all, you are in noticeable pain at 2 units, even if 2 units is not noticeably more painful than 1 unit. The fact is that in moving me from 1 unit to 2 units, you put me into noticeable pain, even if you did not make me noticeably worse off than 1 unit. I notice my pain state even if I do not notice the transition.

As I've said a few times, the confusion here is going to perpetuate without some formal representation. More strictly, on one interpretation, what is happening is that you are moving from a state that is indefinitely, definitely painful to a state that is definitely, definitely painful. So, the person in the former state is on a second-order borderline of painful states. But the latter person is in pain and on no borderline of painful states. He would say that the latter is clearly painful and the former is something like painful, but not clearly so. On another interpretation, you have someone moving from def. indef. painful to a state that is indef., indef painful. The person in the former state that is reasonably called non-painful to a state that is on the borderline of indefinite pain.

So as you make the case more precise, it brings out how clumsy it is to descibe the case in terms of noticeable pain. There are all sorts of degrees of noticability.

The fact is that in moving me from 1 unit to 2 units, you put me into noticeable pain, even if you did not make me noticeably worse off than 1 unit.

And this is exactly why I think that moving one from i to i + 1 units makes one worse off. It makes one worse off by moving one into noticeable pain and it moves one into noticeable pain (except for the move from 0 to 1 unit) even though one does not make one noticeably worse off than one is at i units. Are you suggesting that one is worse off only if one is noticeably worse off? If so, that's what I'm denying. You can make a person worse off by causing some noticable pain. That's bad. That seems to be to be a truism.

As I've said a few times, the confusion here is going to perpetuate without some formal representation.

I think the confusion will get more confusing with a formal representation that is more complex than the case it is supposed to model. And it is intuitions about cases like these that make a representation plausible, not the other way around. That's why I stick to the case. We can judge the case without taking a stand on very intricate issues regarding the best way to model cases in general.

Are you suggesting that one is worse off only if one is noticeably worse off?

When it comes to pain, yes, of course I'm suggesting it. I'm insisting on it, in fact.

I can either:

(A) bring John into existence at -100 units of pain, and only -100 units of pain. He will not undergo any change in pain levels. He will then shortly thereafter go out of existence.

or,

(B) bring John into existence at +100 units of pleasure, and only +100 units of pleasure. He will not undergo any change in pleasure levels. He will then shortly thereafter go out of existence.

I make John worse off by choosing (A) because I cause him pain. But he is not noticeably worse off. Had I chosen (B) then John would have been in a pleasure state and only a pleasure state. He would have been noticeably better off. In neither case could he be noticeably better or worse off as he could be in one and only one state.

Insert that all important 'not' into "He would...have been noticeably better off."

What do you mean he is not noticeably better off? He is noticeably better off in A than he would have been in B. And very likely he is noticeably worse off than he could have been under another option. What you're after is not this, but the Handicapped Kid case, which is a difficult case. But Handicapped Kid (pace Josh Parsons, http://www.springerlink.com/content/pvu1x3g254q17x45/) depends on controversial metaphysical assumptions most of which I reject.

That Parfit case is a cool one, but my case doesn't involve non-identity problems. Jon is the same in both options.

The point is that John is clearly better off in (B). We can judge this, and correctly, without asking: Can John notice that he is better off? Or, Is it noticeable to John that he is better off? The case is supposed to involve a stipulation to the effect that he can notice a pain and not be able to notice he is worse off in (A). If I can't stipulate this, then you must have in mind some conception of "noticeability" I haven't figured out yet.

I agree it's odd to say that one can have two pains, one a worse pain than the other, where one cannot tell that one is a worse pain than the other. For if one cannot tell, in what sense could it possibly be a worse pain? Something is a pain only if it is noticed as such. Similarly, something is a worse pain than another only if it is noticed as such.

That all sounds fine to me. I hope you don't think I've denied any of that.

I actually think you have denied that, but maybe I've been misreading you. My view might also be misunderstood, I think. What I am defending does not rely on John being noticeably worse off or noticeably better off. I've been saying this:

P. It is possible that 2 units is not noticeably worse than 1 unit, 2 units is noticeably painful and 1 unit is not.

1. The reason you can bring about B is because it is noticeably pleasurable.

2.The reason you cannot bring about B is because it is noticeably painful.

3. The reason you cannot move John from 1 unit to 2 units is because 2 units is noticeably painful and 1 unit is not.

4.The reason you can move John from 0 units to 1 unit is becuase neither is noticeably painful.

Perhaps we have talked acrossed purposes. It wouldn't be the first.

I agree to (1) and (2). I disagree with (3) and (4), sort of. I want to be careful to distinguish between:

(A) X is a pain and Y is a pain and X is a worse pain than Y.
(B) X is a pain and Y is a pain and X is worse for S than Y.

I didn't mean to, and hopefully I have not, asserted that (A) could be true while X is not noticeably a worse pain than Y. I'm open to the idea that one pain can be a worse pain only if that pain is a noticeably worse pain.

I meant to, and hopefully I have, asserted that X can be a pain and a pain that is worse for S than some pain that is not noticeably different from it.

Thus, I mean to distinguish between pains that are worse pains than others, and pains that, though they are not worse pains, are such that they can make their bearers worse off.

I deny (3). That is, I believe there are other reasons not to move John from 1 unit to 2 units. I accept that one reason is that the move is from an unnoticeable pain to a noticeable pain. But I think there is an additional reason. Moving John from 1 unit to 2 units makes John worse off. It makes John worse off even if it does not make John noticeably worse off by causing him some pain that is noticeably different from another pain.

I deny (4). I believe we have a reason not to move John from 0 units to 1 unit. I claim that, even though this move doesn't make John noticeably worse off by casuing him a pain that is a noticable different pain, it makes him worse off by causing him a pain.