Does it violate free will if the person was never given the desire to do something in the first place?

Great Debates
441

But the problem is that you are claiming to know something, namely, that the probability of the evidence (seeing evil, seeing orange) given the truth of your hypothesis (tri-omni God, orange-hating John) is smaller than the probability of seeing that evidence given the falsity of the hypothesis. This is mathematically equivalent to saying that ‘observing the evidence should lower my belief in the hypothesis’. The problem is, you don’t know that. So you’re implicitly asserting knowledge you don’t have.

I’ll give a short demonstration of this for the curious:

Summary

Let H be the hypothesis, E the evidence, and \neg H denote the falsity of the hypothesis. Bayes rule tells us how to adjust the belief in H upon encountering the evidence E:

P(H|E) = \frac{P(E|H}{P(E)}\cdot P(H)

Now the question is, when is P(H|E)<P(H), i.e when does the evidence allow us to judge the hypothesis less likely? This leads to:

\frac{P(H|E)}{P(H)} = \frac{P(E|H)}{P(E)}<1,

which directly leads to P(E|H)<P(E). So it has to be the case that the truth of the hypothesis lowers the probability of the evidence being observed. That’s already something that isn’t obviously true, but we can make this more intuitive.

By the law of total probability, we have that

P(E) = P(E|H)\cdot P(H) + P(E|\neg H)\cdot P(\neg H).

For this to be larger than P(E|H), it suffices that

P(E|H) < P(E|\neg H),

because then, any addition of P(E|\neg H) weighted with some positive P(\neg H) will make the result bigger than P(E|H).

Consequently, only if P(E|H) < P(E|\neg H) is it rational to lower belief in the hypothesis upon observing the evidence.

If that were true, that is if the expected value of orange were exactly zero, then in fact any observation of orange would directly make the hypothesis that John abhors orange logically false, because this entails that the probability of seeing orange given that John abhors orange is exactly zero. If there is any positive probability of seeing an orange item on John even though he abhors orange, then the expected amount of orange seen on him is likewise greater than zero.

In this as well as in the case of the PoE, we know that there is a positive expectation of seeing the respective kind of evidence because we know the intersection of cases where the hypothesis is true with cases where we see the given evidence is non-empty.

I’m being perfectly consistent: we have no reason to believe any amount of evil to be inconsistent with the existence of God, as long as we have no information on the amount of evil to expect in a world with a tri-omni God (except that it is non-zero). The arguments that certain instances of evil are gratuitous give us exactly this information, and to the degree that they are inductively convincing, reduce the probability of the truth of the hypothesis (i.e. God’s existence). It yields exactly the information that is needed: something we are less likely to observe given the hypothesis of a tri-omni God than we would be given its falsity.

442

You’re right, but I should say there was a typo in that post you quoted.
I meant to say “any reason whatsoever to doubt that an omnimax God exists?”

Because this is the absurd bullet that was bitten back in post #220 – that no observation of suffering gives us reason to doubt that there’s an all-powerful, all-loving god because there’s a hypothetical universe where these things are compatible.

A line of reasoning that could be used to defend any claim, no matter how unlikely or absurd.

443

Correct. Of course I expect the probability of seeing John in orange to be lower than a random person, if the claim were true, all else being equal.

Do you think the probability of seeing someone in blue, given the claim that they adore blue, and seek to maximize blue in their surroundings were true, is higher than a random person? Well John’s hatred of orange could be rephrased as maximizing all colors that aren’t orange.

No, you misunderstood me here – I’m not saying it’s your fault, this was just an ambiguous sentence in English.

I am saying that we don’t have an expected value of orange. Not that we expect it to be zero.
We are not setting any line or expectation. Merely that the more orange we see, the more we doubt the claim.

And that right there is the flaw in the logic. Well, one of them. We do not have any reason to expect non-zero orange or non-zero suffering.
The fact that a universe with omnimax God and suffering is not known to be impossible tells us nothing because:

  1. That hasn’t proven that it is possible. As I said upthread, if I refute an argument against my dark matter model, it may yet be impossible for other reasons.
  2. It’s adds nothing to probability. Because the number of possible worlds is infinite (especially since we are adding supernatural agents). The p value of any given world, sans evidence, is zero. This is why we don’t use “set of worlds” reasoning in real life – you can almost never conclude anything.
444

I quoted that Aquinas passage too. The antecedent is God’s will: he is saying if God’s will was to do evil, God would do evil, and the proposition as a whole is true, but he denies that God’s will can ever be to do evil (for that would imply privation).

~Max

445

Then you’re both bringing in additional data (i.e. the information you gathered by sampling the amount of orange on lots of other persons; recall in the PoE case, you only get a single world to investigate), and making unwarranted assumptions (that John should be expected to be free in his selection of colors). As I said, given such assumptions, it’s an entirely reasonable case, but without them, you’re just committing a logical error.

And this is really the crux of the problem: you keep accidentally smuggling unwarranted background assumptions in that may be reasonable in everyday cases, but whose heuristics fail to apply for more general situations. Here, it’s that you generally expect people to have a high degree of control over the colors of their possessions, which is entirely reasonable; but it’s not true in general, and for God, it may be entirely true that there is no possibility of getting rid of more evil. You’re applying intuitive reasoning to a case where you have no relevant intuition to appeal to, and thus miss how your statements fail to generalize. That’s why I prefer some rather more contrived, but also more easily controlled scenarios, as in the raisin analogy: there, it is absolutely plain that you don’t have the information necessary to decide whether a raisin cookie should negatively impact our belief.

Great, because that is exactly what it means to have no further information. Then, it’s exactly the case that you don’t know whether evidence should increase or decrease your confidence in the hypothesis, because if you know nothing about the expected value, you know nothing about the probability with which we should expect to find evil, orange, or raisins.

So you’re saying you disagree with the FWD now? Because that we have a non-zero expectation of evil is exactly its conclusion. If the FWD is apt, then there are worlds in which God exists, and evil does, too, which means there is a finite probability for there to be evil in a world in which God exists, which means there is a non-zero (but unknown) expected value of evil in a world where God exists. Those are just all logically equivalent statements.

Again, you’re still getting this the wrong way round. It isn’t merely not known to be impossible, it’s known to be possible.

446

Agreed, but there is a difference in the notion of omnipotence depending on whether the counterfactual is true. @Voyager, as far as I understand, claims that omnibenevolence is such that God would be absolutely prevented from doing evil, even if it were their will to do so. I think an entity which, were their will such as to do evil, could, is entirely as omnipotent as one that actually does evil, because there is no constraint on their powers.

447

The problem is in the inductive proposition - it is disanalogous to most theist’s belief in God.

To wit, if I know John hates orange because I am John and I know what I hate, then it does not matter how much orange I see in John’s world.

~Max

448

That’s just another example of trying to “defend” God by postulating that he’s far, far less powerful than we are. A helpless God is not a tri-omni God, however.

Arguments like this don’t work because they directly contradict observed reality. We know they are false. And no, labeling observation of the physical world “intuition” doesn’t make it true.

449

And as usually, you’re cordially invited to consider the actual argument rather than merely making bland declarations.

450

Best as I can tell, your original example with John can be rephrased:

If an agent repeatedly engages with X voluntarily, then it is unlikely that the agent strongly dislikes X (all else equal).

But you suggested that even if John doesn’t voluntarily put on the orange clothes, that is still evidence against his hatred of the color. So the less agreeable:

If an agent repeatedly engages with X, then it is unlikely that the agent strongly dislikes X (all else equal).

But this assumes agency, a property that doesn’t necessarily extend to God given Plantinga’s free will defense. At least not with the same connotations that give the proposition its sense. Recall that Plantinga denies compatibilism, yet assumes God must act omnibenevolently as a matter of necessity. In the alternative, all else may not be equal for God vis a vis John if the instrumentality of engaging with X is disanalogous. Perhaps most importantly, engagement is disanalogous. God engages with literally everything if you assume God sustains all creation.

~Max

451

I did consider it and argued against it, throwing out insults doesn’t change that. And an insult is not a counterargument, it’s an evasion.

452

None of them were aware of the reason behind natural disasters, or of ways the world could have been designed to prevent them. They might just as well justify God from the placement of the stars in the actual firmament. (It wouldn’t surprise me if they did.)
And I don’t get how natural evil stems from the knowledge of good and evil. Did God create plate tectonics after the fall? Would they not realize that the supposed children of the unfallen couple dying in an earthquake is bad? I’d read that stuff, but it would come after reading some of the Talmud which comes after reading the unread books in my library, so I’d not get to it until long after I find out for myself - or if I live to 120 or so.

453

Doing evil is only logically impossible if omnipotence and omnibenevolence come only as a package. It is possible for an entity which is only omnipotent, so doing evil is not impossible for an omnipotent being per se. That’s the reason it is logically impossible for a being to be both.
This has nothing to do with circularity. The circularity I mentioned was justifying omnibenevolence in the light of evidence against it by claiming that an action by god is good because god is omnibenevolent, though we may not know how it is good. That only is a plausible argument for prescriptive omnibenevolence. You can’t really use it for descriptive omnibenevolence, though I suspect it would be easy to retreat back to it.

454

Not quite. An omnipotent being could do evil, but he would then no long be omnibenevolent, in the prescriptive sense. (Or, more accurately, never was.) If a prescriptively omnibenevolent being could not do evil, due to its nature, it would not be omnipotent. If it is descriptively omnibenevolent - doing no evil up to this point - no problem. It becomes almost omnibenevolent, and the description changes slightly, but I’m not sure it is logically impossible.

455

Geez a lot to reply to. Here we go:

Take out the random person then. All that’s needed for the analogy is that the person has control over their environment.

Omnimax God has absolute control, by definition. It is not me smuggling in an assumption, it is you throwing in claims that e.g. we know there must be a non-zero amount of suffering. We don’t.

Nope. Whether observations support a claim does not require setting some “anticipated level”.
I don’t need to say in advance what level of suffering I expect an omnimax god would allow – we only know by definition that his goal is zero. We can speculate about mitigating factors, but we don’t know that there are any. Any amount (and there’s a fuckload of natural suffering in the world, so this is all a bit silly) supports the claim that he’s not omnimax or doesn’t exist.

And again, before anyone says it: “supports the claim” doesn’t necessarily mean conclusive. This kind of reasoning from observation is never certain, but just as I can reason that my neighbour whose entire home, car, clothes are emblazened with manchester united badges probably likes manchester united, so I can reason that a god with absolute power over this universe presiding over a fuckload of suffering probably isn’t perfectly loving (or doesn’t exist).

If you mean Plantinga’s argument, I don’t think it supports your claim about it. As evidenced by the cites I gave earlier – and it hasn’t gone unnoticed that on 3 times of asking you haven’t produced a cite that Plantinga’s argument defeats the evidential problem of evil.

No it isn’t. The default skeptical position is that we neither know something to be possible, nor impossible. The logical problem of evil suggests that the existence of an omnimax god and suffering is an impossible state of affairs, and plantinga’s argument defeats that. Fine – it’s shown that one argument for that state of affairs being impossible failed.
That is not the same thing as proving that it is possible.

If you want to say as far as we know it’s possible, then fine, but of course this cannot be used for the kinds of inferences you are trying to make.

456

It’s analogous to the premises of this problem. If we’re taking the position that I think you’re alluding to again – of suffering being a good in itself – then the problem of evil basically doesn’t exist in the first place.

My response to that line of argument would be to point out the extremely random and often unfair distribution of suffering. And how suffering persists even when the person can do nothing about it – what’s the benefit to the pain suffered in terminal cancer for example?
And if the response to that is to shrug and suggest it makes sense from a god’s point of view, then we’re back to “mysterious ways”.

If anyone has agency, it’s omnimax god.
We can’t use the hypothesis that external factors might tie god’s hands, to then set as a premise that god’s hands are tied. We don’t know that anything limits god. We don’t know that he couldn’t make a universe with free will and no suffering.

457

Which is exactly what you don’t know, but assume anyway.

Again, if we didn’t know that, the logical PoE would work, and any amount of evil at all would conclusively disprove the existence of God. Either the logical PoE works, or we should expect a non-zero amount of evil, because it is exactly the fact that God and evil can coexist that both defeats the logical PoE and means that there is a non-zero expected amount of evil.

I have explicitly demonstrated how differences in the expected amount lead to a different effect of evidence. You haven’t offered any counterargument beyond blanket denials like this.

But the question isn’t what the goal is, but whether that goal can be attained. And we now it can’t always—again, otherwise, the logical PoE would work. We don’t know to what degree it can. But that degree is exactly what determines the effect of the evidence we have.

Take again the basic logical picture (which you have so far ignored). Two propositions being compatible means that you can’t appeal to one as evidence against the other. The structure of the argument you’re proposing is:

  1. There are things that are both square and red.
  2. This is a red thing.
  3. Therefore, it probably is not a square.

This is, I trust you’ll agree, clearly invalid. But it’s exactly what you’re proposing:

  1. There are worlds in which there is both a tri-omni God and evil.
  2. This world contains evil.
  3. Therefore, there probably is no God.

Now, of course you’re going to say, but God is opposed to evil, so the analogy would only work if squares were opposed to being red. But that doesn’t do anything: the question, in the end, is whether squares have any say in the matter. It’s outcomes that matter, not intentions. In the end, it’s just a numbers game: does God, or the squares, or whatever, ultimately have the power to make their will manifest in a relevantly increased fraction of cases? And the only answer we have at this point is that we don’t know. We know they can’t always do so, but we have no information on whether they almost always or almost never get their wish. Any stipulation to the contrary is an unwarranted assumption without further information. (Again, that already follows from the simple fact that the two propositions are consistent. But for some reason, this never seems to get through to you.)

Take again the visualization I provided:

The FWD ensures that the area of G\cap E is non-zero. Each of the worlds in this overlap has some certain amount of evil. Therefore, there is a non-zero expectation for a certain quantity of evil within any world in the set G, because this just amounts to the average over all the elements of this set, and will be positive as long as some of these elements have some non-zero value of evil. But whether or not the evil in the actual world then should be surprising, given that there is a tri-omni God, depends on that average. And since we don’t know that, we simply don’t have the information to assess this.

This directly contradicts your earlier claim that you believe the FWD works. Because it is only possible for any amount of suffering to cast doubt on God if G\cap E is empty.

I don’t know why you keep banging on about this; I have no problem at all with inductively established conclusions. In fact, all of the calculations I gave are explicitly about assessing the likelihood of a claim, not definitely showing it to be true, so the point you’re defending against here just isn’t one anyone has raised, as far as I can tell. The question is merely whether you can validly decrease your belief in a hypothesis, given the evidence you point to.

As noted, the cites say exactly what I’ve been saying, namely that the FWD means that just pointing to the existence of evil doesn’t suffice to cast doubt on God’s existence. That you just ignore that I’ve pointed this out several times now doesn’t make it any less true.

Because, for the third time now, that’s not a claim anybody is making:

Except, of course, if its possibility has been demonstrated, which it has—that’s the FWD.

And it does so by explicitly showing it to be possible. It constructs a case in which suffering exists despite there being a tri-omni God, which by direct demonstration shows it to be possible for such a case to exist.

Again, that would be false: barring any counter to the FWD, we indeed do know that it’s possible for there to be a world in which God and evil exists. If you accept the FWD, then you accept this, too.

I mean, still, I’m curious: if what you’re saying were right, then evidential arguments would not need to make any reference to gratuitous evil, Rowe would not need to argue for cases of suffering that a tri-omni God both could and would prevent, and so on. Why do you think people go through all this rigamarole? Why do your own cites state, over and over again, that evidential arguments must appeal to evil so great that it can’t plausibly be reconciled with the existence of a tri-omni God, if, as you say, any amount of evil serves to create doubt?

458

This is a controversial claim, rejected by Augustine and virtually all Christians in history, except the Gnostics and Manichaeans.

~Max

459

Maybe so, but I don’t see how.

We are not. You singled out Reformation theologists a few posts back so I dropped the virtue ethics.

So long as there is no contradiction, that is … (Checks thread) a win condition for the libertarian free will theist who assumes the existence of God before confronting the evidentiary problem of evil.

We know that God’s power does not contradict itself. No square circles. And we know that God is omnimax, as you say. These are just part of the nature of God, not external factors (by definition there is nothing external to God).

Therefore if you want to show a contradiction, you must do more than admit ignorance as to whether there is unnecessary suffering. You have to go the extra step and show that there actually is suffering to which an omnibenevolent being would not abide. If you choose to work the evidentiary problem of evil, you have to show that observed suffering implies the existence of unnecessary suffering. This is impossible, in principle (not practice), because belief in God is not inferred from evidence and the existence of God implies nonexistence of unnecessary suffering. Hence: “Then I will also confess to you, That your own right hand can save you.”

If you want to show that observations of suffering provide strong evidence against the existence of God, without assuming the existence of God, that’s a very different debate…

~Max

460

Probably you’re confusing me with someone else, as I haven’t mentioned Reformation anything.

But to re-iterate my point: the description of the problem of evil assumes that evil (I prefer the word “suffering”) is a negative in itself and an omnimax god would want it to be as close to zero as possible. As Hume put it: “Is [God] willing to prevent evil, but not able? then is he impotent. Is he able, but not willing? then is he malevolent…”
Also the OP goes with that premise.

So, if you want to suggest the idea that suffering is a net good in some way, and is a necessary step for the soul or whatever, then that’s fine as a proposed solution to the PoE. But you were suggesting it was an assumption that *I* was making, and I am not; I am just engaging with the argument, which has that as a premise.

I genuinely don’t understand your response.
But yeah, if we’re satisfied with “mysterious ways” then there’s always a “win”, because it’s a universal solution to any objection.

Agreed.

Firstly, this seems like an allusion to the logical problem of evil, which is not the argument I am making or have made at any point in this thread. Necessary suffering might exist, sure.

But secondly, again, I’m engaging with the argument on its terms, I haven’t changed anything. That random acts of death and suffering make people doubt a loving god exists is not something I’ve invented.

My argument, which is basically just restating the evidential problem of evil, is that, all else being equal, we would expect an all-powerful being that wants suffering to be zero to make a world with zero suffering. If there are mitigating factors that mean God must allow N suffering then the burden is on those that propose such factors to define them, explain how they tie an omnimax God and also relate it to the amount of suffering we actually see.