# The Best of all Possible Worlds

**URL:** <https://boards.straightdope.com/t/the-best-of-all-possible-worlds/553190>\
**Category:** Great Debates\
**Created:** [September 9, 2010, 3:46pm UTC](https://boards.straightdope.com/t/the-best-of-all-possible-worlds/553190 "2010-09-09T15:46:04Z")\
**Posts on this page:** 10\
**Page:** 14

<div class="post-metadata">

**Author:** ![begbert2](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/begbert2/32/3357_2.png) [@begbert2](https://boards.straightdope.com/u/begbert2)\
**Post date:** [October 5, 2010, 8:56pm UTC](https://boards.straightdope.com/t/the-best-of-all-possible-worlds/553190/262 "2010-10-05T20:56:03Z")

</div>

> [@Half\_Man\_Half\_Wit](#):
>
> > [@begbert2](#):
> >
> > What do you suppose happens when I _do_ prove that it’s logically possible for it to be the case?
> > 
> > Not that you’re accepting any arguments to that effect. But let’s just play pretend for a moment and say that I _did_ convince you that better worlds are logically possible. What would that do to your argument?
> 
> Well, the argument, as such, would still stand – it would still be possible for god to be both omnimax and for evil to exist in the world. However, if you proved that there exists a world in which there is less evil than in ours, then god obviously wouldn’t actually _be_ omnibenevolent.

This is so illogical I don’t know where to start. In the case where you _agreed_ that god _could_ have made the world better and deliberately _didn’t_, that somehow doesn’t conflict with omnibenevolence until the better world “exists”? And if the better world existed - meaning, that we’d be living in it - then that _wouldn’t_ allow omnibenevolence?

Your thinking here is _literally_ backwards. You’re getting worse the longer you go!

> [@Half\_Man\_Half\_Wit](#):
>
> I have no problem in admitting that. Hell, I believe an omnimax god is incompatible with himself! Of course, one could assume the position that relativity may be necessary for any universe to exist at all, i.e. may be a logical necessity, in which case it would just be impossible for god to know things he logically can’t know – things that are causally removed from him. But that’s besides the point.

Not really - if you accidentally prove ~G as part of your arguments, then you have just guaranteed your failure to find a flaw in the POE. It’s identical to the case of gods that can do the logically impossible in this regard.

> [@Half\_Man\_Half\_Wit](#):
>
> I haven’t claimed anywhere that relativity disproves history, or any such nonsense. Your causal past is set in stone, barring time- or FTL-travel. A global history, i.e. a list of which events follow which, regardless of their occurrence in spacetime, however, is simply not a meaningful concept.
> 
> Thankfully, the problem of evil or relativity isn’t what we’re discussing here.
> 
> No, I’m saying that your world-states are an incoherent concept in a relativistic universe, _and_ that they don’t have anything to do with how possible worlds are commonly understood in valuations of modal logic.
> 
> Nah, relativity’s perfectly fine with modal logic – it’s just your world states that don’t fit.

We may not be discussing the “problem of evil or relativity”, but we’re also not discussing your incorrect model of relativity. Take your objections to world-states (which happen to be _identical_ in function to how possible worlds are commonly understood on valuations of modal logic) and stuff them in the garbage where they belong - because I dismiss all such nonsense as incoherent and incapable of refuting anything I say.

> [@Half\_Man\_Half\_Wit](#):
>
> The problem is, we only know _one_ of those worlds to be logically possible (whichever the one is that we’re in, that is).

Bullshit - anything that’s physically possible is logically possible, necessarily. I’m so sorry that logical impossiblity follows rules and makes logical sense, but unfortunately for your argument, it does.

> [@Half\_Man\_Half\_Wit](#):
>
> This is again just an unsubstantiated assertion. Tell me, what logically possible means would god use to rewind the universe?

He uses f’ing **magic**. He’s **God**. He’s **omnipotent** , which despite your efforts to pretend otherwise is a word that actually has a **meaning**.

You _do_ realize that you can’t break the POE by pointing at styrofoam cups or other nonmipotent gods, right?

> [@Half\_Man\_Half\_Wit](#):
>
> I’m not saying god _can’t_ do anything of the things you allege he can do; I’m merely saying that it’s _possible_ that he can’t, simply since there _are_ things he can’t do. Again, I refer back to the argument I made a couple of posts ago, which you seem to accept.

I seriously doubt I accepted it; I probably just missed it - and it’s not like “a couple of posts ago” is exactly a roadmap to finding it.

Regardless, if you’re talking about your absurd “god can’t make square circles, so that means he can’t make round circles either” argument, then I reject it on account of it being terminally stupid. Your “it’s _possible_ that he can’t” stuff is utterly illogical and completely unsupportable by any valid or sound reasoning.

The logically possible is logically possible. The fact that some other things are not logically possible doesn’t in any way, shape, or form reduce the logical possibility of logically possible things. And this remains true no matter how many times you repeat nonsense to the contrary - even if I miss one or two of the times you do so.

Regardless, I shall lay down the law (again): any god that cannot do the things I listed is a pansy-ass weakling, is _certainly_ not omnipotent (by definition), and is thus is not and _cannot_ be used as any part of a disproof of the POE. You can’t disprove ~G by showing examples of ~Gs.

> [@Half\_Man\_Half\_Wit](#):
>
> Oh, come on, you’re not just gonna pretend I forgot all I wrote in [this post](http://boards.straightdope.com/sdmb/showpost.php?p=12925314&postcount=177), are you? The fact of the matter is that you can leave the ‘G’ part completely implicit, as long as the context is understood. The discussion that’s at hand here is, in a word, whether or not the existence of an omnimax god implies that there can’t be any evil; concentrating on the second part is merely a way to save writing. You did the same thing some posts back:
> 
> > [@begbert2](#):
> >
> > in order to disprove the problem of evil he needs to show that E is true,

1. I didn’t find that incredibly impressive either, but even if it _had_ been, it doesn’t make any stupid and incorrect things you say later any less stupid and incorrect.

2. There is no form of logic in the universe where you can leave the part of the argument you’re trying to prove or disprove “implicit”. Well, aside from bullshit fallacious rhetorical logic, but that’s not a universally recognized system.

3. I left nothing implicit when I said that you need to show E to disprove the POE. Obviously. And that you claim I did shows that you either are bullshitting me, or that you have no idea of how the logic works - and that you have no idea what you yourself are talking about. _You have spent the entire thread arguing that “[]E” is true._ Remember all this nonsense you’ve been saying about your nonmimax god being unable to make a better possible world? About how -in complete defiance of the facts- you are claiming that our actualized world is the only logically possible world? Why do you suppose you’re doing that? Can you remember?

> [@Half\_Man\_Half\_Wit](#):
>
> No. It’s simply that the formalisation of the problem of evil you proposed yields a true statement if there exists a world in which god exists, and there is (whatever amount of) evil (and of course, that world may be ours), a possibility which is precisely antithetical to the logical problem of evil (which asserts that both the existence of an omnimax god and the existence of evil can’t be true).

This is blatantly false. The definition of “omnimax” is logically incompatible with unnecessary evil in any specific world. There may exist possible worlds with a god and no evil, or with evil and no god, or with neither a god _nor_ evil, but at no point have I allowed for the possibility that the set of worlds with both a god and evil is non-empty. Becuase such worlds are impossible. By definition.

If you think I’ve stated that there might be a possible world where an omnibenevolent god chooses evil, then you’ve simply misunderstood what I said. Which wouldn’t surprise me at all at this point.

> [@Half\_Man\_Half\_Wit](#):
>
> I’ve got no idea where you were trying to go with the red balls, incidentally, and what any of that had to do with what you bolded in my post. Basically, the possibility that a ball is red or not does not impinge on whether or not it is actually red; \<\>R is true in any world, whether or not R is, as long as R isn’t necessary. If you mean to say that if the ball in some world is not red, then \<\>R isn’t true (in that world), then that’s just the same error you made earlier, that P -\> P. \<\>R is not a claim about that specific world; it’s a claim about _every_ world.

> [@begbert2](#):
>
> So:  
> in W1, R is true, and ~R is false.  
> in W2, R is false, and ~R is true.
> 
> With me so far?
> 
> So: in this scenario, R is false, ~R is false, ~R is true, ~~R is true, \<\>R is true, \<\>~R is true, ~\<\>R is false, and ~\<\>~R is false (to cover all the bases).

Underlining added - Gee, I _wonder_ if I was trying to say that \<\>R isn’t true. It’s unfortunate that I didn’t make that clear, forcing you to have to guess and put together a blatant strawman to pretend that I don’t understand something that I obviously, obviously do.

However, I’m not sure _you_ understand modal logic; you keep writing about it being **possible** that in **our specific world** there’s evil, when of course in actuality \<\>X is not a claim about any _specific_ world; it’s a claim about the set of _all_ possible worlds.

And incidentally, pointing out this confusion your part was the whole point of the red ball example. Would you care to answer the following question I asked in the prior post? I would find your answer instructive…and so might you:

> [@begbert2](#):
>
> So: R is ‘possible’, in logical vernacular.
> 
> Does that mean that it’s ‘possible’ (english vernacular) for R to be true in W2? By your argument, it does. By your argument you’ve just disproved the statement “in W2, R is false, and ~R is true.”

> [@Half\_Man\_Half\_Wit](#):
>
> The bolded part in my post referred to the inadequacy in your proposed formalization of the POE – that the implication yields a true statement when god isn’t necessary, i.e. when he’s contingent; so that if in some world (ours, for instance), an omnimax god exists, and evil also exists, as long as god doesn’t exist in every possible world, G -\> ~E is a true statement. Which of course means that the existence of evil doesn’t mean that there’s no god (only that there’s _not necessarily_ a god), which is the opposite of what you’d want a POE to say.

Come back when you know how modal logic works. For starters, learn that (\<\>A and \<\>B) does not imply \<\>(A and B).

For a starter on that, consider the case where B is ~A: that is, ask yourself whether (\<\>A and \<\>~A) implies \<\>(A and ~A). As long as you think the answer is “yes”, goto 10.

> [@Half\_Man\_Half\_Wit](#):
>
> If it’s hard to see, just work through the argument: your formalization says, ‘if there is necessarily a god, then there is necessarily no evil’. In true modus tollens form, we observe that there is, in fact, evil, and hence, not necessarily no evil. This means that the antecedent must be false – that there is not necessarily a god. Of course, this means that it’s still possible for there to be a god – in short, the existence of evil means that there possibly is a god – and the argument has gone officially nowhere.

Oh, I understand the argument just fine. I simply _also_ understand that (\<\>G and \<\>E) doesn’t imply \<\>(G and E).

> [@Half\_Man\_Half\_Wit](#):
>
> > [@](#):
> >
> > The POE talks specifically about what’s going on in _our_ world, using observable evidence about the evil in _our_ world as a cite.
> 
> No, in its original form, it asserts the incompatibility of a set of propositions – that god is omniscient, that he is omnipotent, that he is omnibenevolent, and that there exists evil. Of course, this last one is derived from observation; however, this is a claim of logical impossibility, and hence, a claim about impossibility in all conceivable circumstances – taking into account modalities when discussing this is merely natural.

Right. However if one uses modal logic _correctly_, they find that it doesn’t change the results of the argument that lacks it. That’s the thing about logic; 1+1 keeps equalling 2 even if you do the same calculation in a different way.

Do yourself a favor and try to write your argument in _actual formal logic_, not this fallacious rhetorical crap. If you actually did that, and stuck to legal rules of logic, I think you’d find your errors rather quickly.

---

<div class="post-metadata">

**Author:** ![Half\_Man\_Half\_Wit](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/half_man_half_wit/32/21766_2.png) [@Half\_Man\_Half\_Wit](https://boards.straightdope.com/u/Half_Man_Half_Wit)\
**Post date:** [October 6, 2010, 11:37am UTC](https://boards.straightdope.com/t/the-best-of-all-possible-worlds/553190/263 "2010-10-06T11:37:40Z")

</div>

> [@begbert2](#):
>
> This is so illogical I don’t know where to start. In the case where you _agreed_ that god _could_ have made the world better and deliberately _didn’t_, that somehow doesn’t conflict with omnibenevolence until the better world “exists”?

I’m not really sure where the problem is. If god could have put us in a better world, but didn’t (i.e. if you prove that there is a logically possible world in which there is less evil), well then obviously that’s it for omnibenevolence. But Leibniz’ argument works despite that – it still is possible for an omnimax god and a world in which evil exists to coexist; that just isn’t the case in actuality. God could have been omnibenevolent, logically; however, he isn’t in actuality. Like my car could have been white, but actually is black – doesn’t mean it’s not true that it could have been white.

> [@](#):
>
> Not really - if you accidentally prove ~G as part of your arguments, then you have just guaranteed your failure to find a flaw in the POE.

Well, if relativity disproves god, then you should get the word out, seems like the Vatican didn’t get that memo.

The thing is that while I know the world states are relativistically incoherent, I don’t know that there isn’t a way for god to exist ‘outside time’ or something like that; I don’t consider it very likely, but I also can’t dismiss the possibility.

> [@](#):
>
> We may not be discussing the “problem of evil or relativity”, but we’re also not discussing your incorrect model of relativity.

I’d like to see an argument for this incorrectness.

> [@](#):
>
> Take your objections to world-states (which happen to be _identical_ in function to how possible worlds are commonly understood on valuations of modal logic)

Nah, modal logic typically doesn’t cripple itself like that; and if you intend on insisting that it does, I’d like to see a cite.

> [@](#):
>
> I dismiss all such nonsense as incoherent and incapable of refuting anything I say.

You do indeed, problem is you lack any reasonable grounds to do so.

> [@](#):
>
> Bullshit - anything that’s physically possible is logically possible, necessarily.

And Last Thursdayism is physically possible? Not by any physical theory I know.

> [@](#):
>
> He uses f’ing **magic**.

So now you want me to grant you _magic_ as logically possible, based entirely on your assertion?

> [@](#):
>
> He’s **God**. He’s **omnipotent** , which despite your efforts to pretend otherwise is a word that actually has a **meaning**.

Yes, and that meaning is not ‘can do everything that’s necessary for your argument’.

> [@](#):
>
> Regardless, if you’re talking about your absurd “god can’t make square circles, so that means he can’t make round circles either” argument,

That is rather a misrepresentation.

> [@](#):
>
> then I reject it on account of it being terminally stupid. Your “it’s _possible_ that he can’t” stuff is utterly illogical and completely unsupportable by any valid or sound reasoning.

The reasoning in the argument I provided [in this post](http://boards.straightdope.com/sdmb/showpost.php?p=12972061&postcount=251) is perfectly valid by anyone’s standards. It’s elementary – for your convenience: There is something that is an x that is P; hence, it is possible for something that is an x to be P; y is an x; hence, it is possible for y to be P. I don’t really see any room for argument there.

> [@](#):
>
> The logically possible is logically possible. The fact that some other things are not logically possible doesn’t in any way, shape, or form reduce the logical possibility of logically possible things. And this remains true no matter how many times you repeat nonsense to the contrary - even if I miss one or two of the times you do so.

I’ve never asserted the contrary. Perhaps it’s easier if we leave the ‘logically possible’ aside for a moment and concentrate on ‘things that can’t be done’. Those two categories are equivalent, for all we know, but the latter is perhaps less of an intuition pump; while one generally thinks that everything one can imagine is logically possible, that’s perhaps not so for the latter.

So we know that there are some things that can’t be done, right? Hence, one is never justified to a priori assume that any given thing is a thing that can be done. That’s really the core of the argument. Of course, one is similarly unjustified in assuming that any given thing is a thing that _can_ be done, and you keep claiming that this is what I am arguing for; however, it isn’t – I merely argue for the (to my mind, absolutely uncontroversial) assertion that it’s possible for everything to be a thing that can’t be done. It’s likewise possible for everything to be a thing that _can_ be done, and I’m perfectly fine with it – however, I trust you see the asymmetry in our positions by now: for everything you assert to be something that can be done, you carry the burden of proof; I merely state the default, born out exemplary, that for every thing, it may be the case that it can’t be done.

Your task then, to prove that something can be done, seems impossible in principle – it’s Hume’s problem of induction once more: seeing x being done gives you no grounds to conclude that y can be done; indeed, strictly speaking, seeing x being done yesterday gives you no grounds to conclude that x can be done tomorrow – having seen the sun rise yesterday does not mean that the sun will rise tomorrow.

> [@](#):
>
> Regardless, I shall lay down the law (again): any god that cannot do the things I listed is a pansy-ass weakling, is _certainly_ not omnipotent (by definition),

Well, that’s a bit of an easy way out, if you just get to define yourself to be right, don’t you think?

> [@](#):
>
> 1. There is no form of logic in the universe where you can leave the part of the argument you’re trying to prove or disprove “implicit”. Well, aside from bullshit fallacious rhetorical logic, but that’s not a universally recognized system.

I’m not trying to prove or disprove the existence of god; the existence of god is an assumption in the POE, that is brought to an alleged contradiction (if you didn’t assume god, there would be nothing to reconcile with the existence of evil – i.e. if G is false, G -\> ~E is always trivially true.) This assumption is left untouched, so I can indeed leave it implicit – the reason for that being that both G -\> ~E and its negation, G & \<\>E, yield true statements precisely iff their second half is true, provided G is true.

> [@](#):
>
> 1. I left nothing implicit when I said that you need to show E to disprove the POE. Obviously. And that you claim I did shows that you either are bullshitting me, or that you have no idea of how the logic works - and that you have no idea what you yourself are talking about. _You have spent the entire thread arguing that “[]E” is true._

OK, this is something you really finally need to get straight. I have never in this thread argued that E is true, nor will I ever do that. It would be an unnecessarily strong argument to do so.

> [@](#):
>
> Remember all this nonsense you’ve been saying about your nonmimax god being unable to make a better possible world?

_Possibly_ unable. I mean, really, after all this time? Do you just still not get what has to be established to show the POE wrong, or are you just trying to strawman me?

> [@](#):
>
> About how -in complete defiance of the facts- you are claiming that our actualized world is the only logically possible world?

Just out of interest, what ‘facts’ would be contradicted by that? I mean, obviously, we only experience one world, so all our experience is trivially consistent with there only being one possible world.

> [@](#):
>
> This is blatantly false. The definition of “omnimax” is logically incompatible with unnecessary evil in any specific world. There may exist possible worlds with a god and no evil, or with evil and no god, or with neither a god _nor_ evil, but at no point have I allowed for the possibility that the set of worlds with both a god and evil is non-empty. Becuase such worlds are impossible. By definition.

Yes, that is indeed what the POE says; it’s just not what _your_ POE says.

> [@](#):
>
> However, I’m not sure _you_ understand modal logic; you keep writing about it being **possible** that in **our specific world** there’s evil, when of course in actuality \<\>X is not a claim about any _specific_ world; it’s a claim about the set of _all_ possible worlds.

Of which ours is one. And perhaps the only one, at that.

> [@](#):
>
> And incidentally, pointing out this confusion your part was the whole point of the red ball example. Would you care to answer the following question I asked in the prior post? I would find your answer instructive…and so might you:

In W2, \<\>R is true. In W2, R is false. There is no tension between these statements.

> [@](#):
>
> Oh, I understand the argument just fine. I simply _also_ understand that (\<\>G and \<\>E) doesn’t imply \<\>(G and E).

Well that’s… nice I guess, but irrelevant. Your POE yields ~G upon the assertion ~~E. It thus fails at disproving the existence of god; it merely disproves the necessity of his existence, which never was under debate. Sure, it might be that things shake out such that there is no world in which G and E are both true, but that isn’t established by your argument (and is an a priori possibility without making any argument at all) – or only if you assume that G and E are in contradiction with each other, which would be begging the question.

> [@](#):
>
> Right. However if one uses modal logic _correctly_, they find that it doesn’t change the results of the argument that lacks it.

Well, as has been noted in this thread, Alvin Plantinga thinks differently, and a majority of contemporary philosophers agree.

---

<div class="post-metadata">

**Author:** ![begbert2](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/begbert2/32/3357_2.png) [@begbert2](https://boards.straightdope.com/u/begbert2)\
**Post date:** [October 6, 2010, 8:41pm UTC](https://boards.straightdope.com/t/the-best-of-all-possible-worlds/553190/264 "2010-10-06T20:41:06Z")

</div>

**Half Man Half Wit** , a large percentage of your prior post can be answered with a single statement, one which I have been repeating over and over again for most of the thread: you are brutalizing the ambiguous meaning of the word “possible”.

The word “possible” has two separate and distinct meanings, which must not be confused:

1. Logical meaning: that in the set of all possible worlds, there is at least one world in which the statement in question is true.

2. English meaning: in a **given** world, it is not yet known that the statement in question is not true.

For an illustration of this difference, consider the case where you are sitting inside your house, and are suddenly struck with curiousity over whether it’s raining outside, since you don’t happen to know. Specifically, you ask yourself, “Is it possible that it’s raining?”

In the english sense of the word, the answer is “yes”. This is because the questioner hasn’t ruled out the possibility that it’s raining in their possible world yet, due to lack of information about their specific possible world. If you got up and went and looked out your window, though, you would add a few more axioms to your knowledge base/premise set, including either “it’s raining” or “~(it’s raining)”, depending on which it happens to be. If it’s not raining, then , the answer to “Is it possible that it’s raining?” changes to ‘no’; and if it _is_ raining, then the answer stays ‘yes’…but most people wouldn’t put it that way, since we don’t tend to think of known facts as ‘mere’ possibilities, in english. Regardless, in either case the question depends only on the state of things in the single specific world under consideration.

Now, the _logical_ sense of the word: the answer to the question is…well, it kind of depends on what’s happening in the complete set of logical worlds, now doesn’t it? If you happen to _know_ that it’s logically possible for it to be raining, which is to say that you know it’s raining in some logically possible world, then you can confidently assert “yes”. If you don’t know that, then you can’t assert either \<\>(it’s raining) or \<\>~(it’s raining), because you just don’t know either way. And when you look out the window, only _one_ of those questions is answered, and the other isn’t! If it happens to be raining, then you then know \<\>(it’s raining) is true, since you’ve verified that in one possible world it is. But if it’s not raining, then you _don’t_ know if \<\>(it’s raining) is true or not; you can only assert that \<\>~(it’s raining) is true, because the only possible world you’ve verified only gives an example of ~(it’s raining) and not (it’s raining).

Now, the astute person will notice that I just said, if you haven’t looked out your window yet, that rain would be possible in the english sense and might _not_ be possible in the logical sense. And yes, you read that right. This is because the english sense of the word _has nothing to do with logical possibility at all_. What it’s about, is _ignorance_. If you are _ignorant_ of whether something is true or not, then you can say “it’s possible”. It you know either way, then you generally don’t use the term. Of course if you happen to know something is logically impossible then you’re not ignorant anymore and it’s no longer possible in the english sense either. But that doesn’t mean that the english sense is the same as the logical sense; not at all. It simply isn’t; the closest logical analogue to the english sense of the world is not \<\>A, but instead (A or ~A).

As the terms mean completely different things, you need to be very very careful about knowing which one you are using, and being clear about which one you are using. In my opinion the best way to go about this is to very deliberately never use the english sense at _all_. To do otherwise risks fallaciously swapping one with the other. As you do _constantly_.

The point of the red ball scenario was to call to your attention that you are doing these fallacious swaps. Essentially I’m going back to modal logic basics in the attempt to get you back on the correct page. If that seems like I’m insulting your intelligence in doing that, well, what else am I supposed to do?  
Now, on to your post itself. I won’t necessarily respond to everything, just things that I think need further elaboration beyond the above explanation.

> [@Half\_Man\_Half\_Wit](#):
>
> > [@begbert2](#):
> >
> > This is so illogical I don’t know where to start. In the case where you _agreed_ that god _could_ have made the world better and deliberately _didn’t_, that somehow doesn’t conflict with omnibenevolence until the better world “exists”?
> 
> I’m not really sure where the problem is. If god could have put us in a better world, but didn’t (i.e. if you prove that there is a logically possible world in which there is less evil), well then obviously that’s it for omnibenevolence. But Leibniz’ argument works despite that – it still is possible for an omnimax god and a world in which evil exists to coexist; that just isn’t the case in actuality. God could have been omnibenevolent, logically; however, he isn’t in actuality. Like my car could have been white, but actually is black – doesn’t mean it’s not true that it could have been white.

Leibniz’ argument is that there is **not** a logically possible world with less evil in it than our world. He is arguing that all the evil in our world, every speck of it, is logically necessary. If he convinced us that he was correct about this, then that would be an effective counter to most variants of the POE - the main exception being that it is susceptible to the accusation that he’s making the god so non-omnipotent that it doesn’t count as omnimax. (That accusation may sound familiar to you too.)

So, Leibniz’ argument does **not** work ‘despite’ a (english) possibility that there’s a better world than ours; he must _disprove_ that (english) possibility to be able to continue. Or at least, talk his listeners into dismissing it as a (english) possibility - which is probably pretty easy for thiests like himself who have a vested interest in accepting his argument and don’t mind cognitive dissonance. Mysterious Ways are _very_ popular among those types as a way of saying “God can’t do this - yet He’s still omnipotent somehow!”

As for the rest, you are confusing the multiple definitions of “possible”. All of your white car stuff is _explicitly_ based on those errors.

> [@Half\_Man\_Half\_Wit](#):
>
> Well, if relativity disproves god, then you should get the word out, seems like the Vatican didn’t get that memo.

There are an _assload_ of memos the Vatican hasn’t gotten. I think we can agree that that’s not proof of anything.

> [@Half\_Man\_Half\_Wit](#):
>
> The thing is that while I know the world states are relativistically incoherent, I don’t know that there isn’t a way for god to exist ‘outside time’ or something like that; I don’t consider it very likely, but I also can’t dismiss the possibility.

God existing outside of time wouldn’t help. If relativity was as you describe it his omniscience and omnipotence would be abrogated if any part of anything anywhere had relativity, because then that would be some part of God’s domain that he could not know/control.

> [@Half\_Man\_Half\_Wit](#):
>
> I’d like to see an argument for this incorrectness.

You’re specifically arguing that the universe cannot be ended at any point in time, _or started_ at any point in time. That is, you are arguing that it’s impossible for this universe and its timeline to have ever begun. That is, you are arguing that this universe does not and cannot exist.

Any loophole you find that allows time to have started via big bang or other method is a tunnel large enough to drive momentary world-states through.

> [@Half\_Man\_Half\_Wit](#):
>
> Nah, modal logic typically doesn’t cripple itself like that; and if you intend on insisting that it does, I’d like to see a cite.
> 
> You do indeed, problem is you lack any reasonable grounds to do so.

Perhaps you’re unaware of what modal logic _is_. It divides the logical sphere into discrete ‘slices’ in which all the logical statements under consideration have a single difinitive True or False value. Like, for example, “Socrates is dead” has a true or false value. Of course, the state of the world changes as time passes; the truth values of statements change. So, when you have statements like “Socrates is dead” (or “God exists”), then you **must** consider the world under consideration to refer to a specific moment in time; what I am calling a “world state”. The possibility of world-states is built **directly** into modal logic; and very arguably (as I’ve said) in most cases the possible worlds you’re talking about are actually possible world-states.

B-b-b-but that’s not right!, you’ll protest. That statement was supposed to be 'Socrates was dead on 10/6/2010" instead! The truth statments have to have time stamps!

That is, of course _one_ way of defining your statements and modal worlds. But it’s definitely not the only way. And there’s no hint of that being the way it’s defined in the POE, either; I don’t see any timestamp on “God exists”.

So yeah - the definition of modal logic is my cite, and the reason I dismiss your denial of possible world-states.

> [@Half\_Man\_Half\_Wit](#):
>
> And Last Thursdayism is physically possible? Not by any physical theory I know.
> 
> So now you want me to grant you _magic_ as logically possible, based entirely on your assertion?
> 
> Yes, and that meaning is not ‘can do everything that’s necessary for your argument’.

The funny/ironic/sad thing is, that’s how **you** are defining it. For the sake of your argument you are defining God to be **absolutely impotent** regarding our universe - literally! You are saying he is unable to break natural law, unable to do miracles, unable to change anything at all in any possible way. And then you have the _audacity_ to claim that that doesn’t stop him from being omnipotent.

I say, hell no. I declare that any omnipotent god has at least as much control over our universe as the writer of a book has over its contents. They can reach in and twiddle, change things, rewrite whole sections, make things happen for no reason, make things appear and disappear, deus ex machina things left and right - _and_, start the book with whatever setting they choose. Including the following setting: modern earth, last thursday.

If you don’t like that, then will all due respect, too fucking bad. Go whine to somebody who cares. The POE isn’t interested in disproving styrofoam cups and it’s not interested in disproving your nonmipotent gods. It’s interested in disproving the gods that people _actually_ call omnipotent, or all-powerful, or capable of miracles, or able to part seas, or able to turn water into wine. Author gods. Gods with power. Gods who can do with the universe as they will.

As long as you restrict yourself to talking about powerless little pansies, the POE in unfazed by your efforts; you are rather literally shooting in a different direction from where your proposed target is standing.

> [@Half\_Man\_Half\_Wit](#):
>
> That is rather a misrepresentation.
> 
> The reasoning in the argument I provided [in this post](http://boards.straightdope.com/sdmb/showpost.php?p=12972061&postcount=251) is perfectly valid by anyone’s standards. It’s elementary – for your convenience: There is something that is an x that is P; hence, it is possible for something that is an x to be P; y is an x; hence, it is possible for y to be P. I don’t really see any room for argument there.

This does not work if you use “possible” correctly. Specifically, this says nothing at all about about whether y is a P in _our_ possible world. Which is the issue under discussion.

> [@Half\_Man\_Half\_Wit](#):
>
> I’ve never asserted the contrary. Perhaps it’s easier if we leave the ‘logically possible’ aside for a moment and concentrate on ‘things that can’t be done’. Those two categories are equivalent, for all we know, but the latter is perhaps less of an intuition pump; while one generally thinks that everything one can imagine is logically possible, that’s perhaps not so for the latter.
> 
> So we know that there are some things that can’t be done, right? Hence, one is never justified to a priori assume that any given thing is a thing that can be done. That’s really the core of the argument. Of course, one is similarly unjustified in assuming that any given thing is a thing that _can_ be done, and you keep claiming that this is what I am arguing for; however, it isn’t – I merely argue for the (to my mind, absolutely uncontroversial) assertion that it’s possible for everything to be a thing that can’t be done. It’s likewise possible for everything to be a thing that _can_ be done, and I’m perfectly fine with it – however, I trust you see the asymmetry in our positions by now: for everything you assert to be something that can be done, you carry the burden of proof; I merely state the default, born out exemplary, that for every thing, it may be the case that it can’t be done.
> 
> Your task then, to prove that something can be done, seems impossible in principle – it’s Hume’s problem of induction once more: seeing x being done gives you no grounds to conclude that y can be done; indeed, strictly speaking, seeing x being done yesterday gives you no grounds to conclude that x can be done tomorrow – having seen the sun rise yesterday does not mean that the sun will rise tomorrow.

You are _grossly_ misusing “possible”. And you cannot handwave your claim that virtually nothing is physically (and thus, logically) possible. This is not seriously disputable; the laws of physics have variables in them and everything, allowing variance in the way things can physically be. _Wide_ variance.

Plus, of course, you don’t get to redefine “omnipotence” to the degree necessary for you to handwave evil into logical impossiblity. Both the physics of reality _and_ the definitions in question disagree with you; are you going to try and reject both simultaneously?

> [@Half\_Man\_Half\_Wit](#):
>
> I’m not trying to prove or disprove the existence of god; the existence of god is an assumption in the POE, that is brought to an alleged contradiction (if you didn’t assume god, there would be nothing to reconcile with the existence of evil – i.e. if G is false, G -\> ~E is always trivially true.) This assumption is left untouched, so I can indeed leave it implicit – the reason for that being that both G -\> ~E and its negation, G & \<\>E, yield true statements precisely iff their second half is true, provided G is true.

You don’t get unstated assumptions in logic. Period.

That kind of shit is only done in rhetorical ‘argument’ - since it’s a breeding ground for fallacy. I do not do that, I am not going to do that, and I am not going to let you do that. If you _try_ to do that, I will call what you write the fallacious gibberish it is.

And you apparently also don’t understand how proofs by contradiction work, either. The existence of god is **very explicitly** what in the POE is attempting to prove/disprove, regardless of how the formulation you’re looking at happens to be laid out. (Not all of them are written as proofs by contradiction, either.)

> [@Half\_Man\_Half\_Wit](#):
>
> OK, this is something you really finally need to get straight. I have never in this thread argued that E is true, nor will I ever do that. It would be an unnecessarily strong argument to do so.

Then you don’t know what the hell you’re talking about. To disprove the POE with a Leibniz-like argument, you **must** convince your audience to accept the premise E - or more accurately, that **all** evil we see on earth is necessary, and thus that there is _no_ unnecessary evil in our world.

Just proving \<\>E gives you nothing at all. As should be obvious, because glancing around at our world proves \<\>E to be obviously true, and that didn’t ever slow anybody making the POE down any.

Your belief that you don’t have to prove E to make a Leibniz-like argument is based on your errors with the word “possible”. You’re wrong.

> [@Half\_Man\_Half\_Wit](#):
>
> Yes, that is indeed what the POE says; it’s just not what _your_ POE says.

I know exactly what I’m saying, thanks very much. If you disagree with me about what I’m saying, that’s _your_ error.

> [@Half\_Man\_Half\_Wit](#):
>
> In W2, \<\>R is true. In W2, R is false. There is no tension between these statements.

Actually I’m not sure that “In W2, \<\>R is true” makes _sense_. That is, I think it’s _explicitly_ gibberish, of the “yellow addition” variety. The way I was taught \<\> only has meaning when referring to a set of possible worlds. I mean, what do _you_ imagine “In W2, \<\>R” means?

> [@Half\_Man\_Half\_Wit](#):
>
> Well that’s… nice I guess, but irrelevant. Your POE yields ~G upon the assertion ~~E. It thus fails at disproving the existence of god; it merely disproves the necessity of his existence, which never was under debate. Sure, it might be that things shake out such that there is no world in which G and E are both true, but that isn’t established by your argument (and is an a priori possibility without making any argument at all) – or only if you assume that G and E are in contradiction with each other, which would be begging the question.

I don’t know what you imagine the POE says, but this isn’t it. And my comment was directly relevent to the blatant error you were making, Try reading it with the text it was responding to in mind and perhaps it will make more sense to you.

> [@Half\_Man\_Half\_Wit](#):
>
> > [@](#):
> >
> > Right. However if one uses modal logic _correctly_, they find that it doesn’t change the results of the argument that lacks it.
> 
> Well, as has been noted in this thread, Alvin Plantinga thinks differently, and a majority of contemporary philosophers agree.

And as I responded to that note in this thread, the Plantinga argument is about libertarian free will and is readily countered based on the fact there’s no such thing. And argumentum ad populum and appeals to authority are fallacies, incendentally.

---

<div class="post-metadata">

**Author:** ![Half\_Man\_Half\_Wit](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/half_man_half_wit/32/21766_2.png) [@Half\_Man\_Half\_Wit](https://boards.straightdope.com/u/Half_Man_Half_Wit)\
**Post date:** [October 7, 2010, 12:58pm UTC](https://boards.straightdope.com/t/the-best-of-all-possible-worlds/553190/265 "2010-10-07T12:58:13Z")

</div>

> [@begbert2](#):
>
> The word “possible” has two separate and distinct meanings, which must not be confused:
> 
> 1. Logical meaning: that in the set of all possible worlds, there is at least one world in which the statement in question is true.
> 
> 2. English meaning: in a **given** world, it is not yet known that the statement in question is not true.

Ah, I think I see what you mean now. What you call the English meaning, I would call the epistemic meaning of possibility; your logical meaning I know as the metaphysical meaning.

However, both are valid logical meanings of ‘possibility’, epistemic (modal) logic being just as much of a logic as alethic modal logics; also, note that I can use possible in English for both meanings, as well: ‘it is possible that black swans exist’ can mean either ‘for all I know, there are black swans’ (epistemic meaning) or ‘whether or not there are any, black swans are possible (in the sense of not necessarily nonexistent)’ (metaphysical meaning).

You accuse me of confusing epistemic and metaphysical possibility. I don’t think I do, but let’s work through this.

You’re right in noting that in your example, there is a difference whether we interpret the modalities epistemically or metaphysically. In an epistemic interpretation, \<\>r (r for _it’s raining_) and \<\>~r are true a priori, since before I look, I don’t know whether it rains or not. After I’ve looked, it is no longer epistemically possible that I don’t know whether or not it rains; if it rains, I know that it rains, hence, r.

That’s not the case metaphysically. After I look out of the window, the knowledge I gain is that r is true (if it rains, that is). I also gain the knowledge that ~r is false. You claim that I gain the knowledge that \<\>r when I look out of the window; that’s not entirely true, however, since it also may be that r (this of course entails \<\>r, but it’s stronger, since it also entails ~\<\>~r, which is not decided by seeing it rain outside). It is precisely true if (and only if) it was true before I looked out of the window.

This is rather intuitive if you think about it – looking out of the window, i.e. gaining epistemic knowledge, can’t change which set of worlds we actually are in, it can only change our knowledge of which sets we may possibly be in.

So, a priori, we are in one of three world-sets:  
A) in which it rains in all worlds  
B) in which it rains in some worlds (and consequently, doesn’t rain in others)  
C) in which it rains in no world.

Epistemically, I can thus assert a priori that \<\>r, or that \<\>~r; both are consistent with my knowledge. Metaphysically, it may be the case that either is wrong, if either r or ~r. You’re right in noticing the need to distinguish between both possibilities.

Now, if I look out of the window, I gain epistemic knowledge, and am able to exclude one set of worlds from the three, either that in which r, or that in which ~r. I am not, however, able to decide which set of worlds I am in precisely. This is, I think, where you’re mistaken.

Let’s look at the white car once more. You going to check its actual colour tells you (epistemically) that we’re in one of two world sets: that in which cars must be its actual colour, or that in which cars may be its actual colour; and it excludes the (epistemic) possibility that we are in a set of worlds in which the car can’t be its actual colour, since, well, it is its actual colour.

Metaphysically, we can (‘actually’) be in either of these two world sets, which means that epistemically, it is possible that we are in the set of worlds in which cars can be a colour different from the one my car does actually have. This is the possibility my argument is intended to establish.

You are thus right in saying that my argument mixes epistemic and metaphysical modalities, however, it doesn’t do so fallaciously. The metaphysical part is in establishing that there exists a world set such that it is possible for both omnimax god to exist, and for there to be evil; and the epistemic part consists of noting that it is possible (epistemically – we don’t know that we don’t) for us to exist in this set of worlds.

> [@](#):
>
> It simply isn’t; the closest logical analogue to the english sense of the world is not \<\>A, but instead (A or ~A).

As I noted, in epistemic logic, the general reading of what you called the ‘English sense’ of possibility is ‘for all x knows, it may be true that’.

> [@](#):
>
> So, Leibniz’ argument does **not** work ‘despite’ a (english) possibility that there’s a better world than ours; he must _disprove_ that (english) possibility to be able to continue.

Leibniz’ argument establishes the metaphysical possibility of god and evil coexisting; if you prove the existence of a better world, you disprove the epistemic possibility of us living in a world set in which this metaphysical possibility is actual. Doesn’t mean that such a world set is no longer metaphysically possible any more than the observation that it’s raining means that it is metaphysically impossible that it doesn’t rain; it is only epistemically impossible.

> [@](#):
>
> God existing outside of time wouldn’t help. If relativity was as you describe it his omniscience and omnipotence would be abrogated if any part of anything anywhere had relativity, because then that would be some part of God’s domain that he could not know/control.

Even this can be gotten around by the hypothesis that relativity is logically necessary, which is a (epistemic) possibility.

> [@](#):
>
> You’re specifically arguing that the universe cannot be ended at any point in time, _or started_ at any point in time.

Well, yes, that’s how current cosmology works – the universe wasn’t created in time, but _with_ time.

> [@](#):
>
> That is, you are arguing that it’s impossible for this universe and its timeline to have ever begun. That is, you are arguing that this universe does not and cannot exist.

That, of course, doesn’t follow.

> [@](#):
>
> Any loophole you find that allows time to have started via big bang or other method is a tunnel large enough to drive momentary world-states through.

They’d still be inconsistent with relativity, though.

> [@](#):
>
> Perhaps you’re unaware of what modal logic _is_. It divides the logical sphere into discrete ‘slices’ in which all the logical statements under consideration have a single difinitive True or False value. Like, for example, “Socrates is dead” has a true or false value.

That’s the problem! In a world state, such as, for example, my spatial hypersurface of simultaneity, there _is_ no Socrates – that word is a non-denoting name, as much as ‘the present king of France’. So there is _no_ definite truth value to the statement ‘Socrates is dead’, other than on those world slices in which it is wrong. And yet, typically, every possible world should be able to assign a truth value to this statement.

The problem relativity introduces is then that my world state, my hypersurface of simultaneity, possibly includes an Andromedan to whom Socrates is alive in _his_ plane of simultaneity (even though he has no way of knowing that).

> [@](#):
>
> B-b-b-but that’s not right!, you’ll protest. That statement was supposed to be 'Socrates was dead on 10/6/2010" instead! The truth statments have to have time stamps!

Well, you always have to include context in any proposition that is dependent on context, of course. Consider one of your world states, and the statement ‘gravity is about g/6’. On its own, it’s meaningless – it needs a ‘space stamp’, like, for instance, ‘gravity is g/6 _on the moon_’. Only then can one assign a truth value to this statement.

> [@](#):
>
> That is, of course _one_ way of defining your statements and modal worlds. But it’s definitely not the only way. And there’s no hint of that being the way it’s defined in the POE, either; I don’t see any timestamp on “God exists”.

Well, just to point that out, I did in the OP refer explicitly to histories in total, rather than timeslices.

> [@](#):
>
> So yeah - the definition of modal logic is my cite, and the reason I dismiss your denial of possible world-states.

It’s just that, in your modal logic, the vast majority of statements are undecidable – like ‘Socrates is dead’, for every world state in which it isn’t false.

> [@](#):
>
> You are saying he is unable to break natural law, unable to do miracles, unable to change anything at all in any possible way.

I’m not sure why I keep trying to explain this, or why you seem to have such a hard time with it – I am not saying that he is unable to do anything at all. I’m saying that for every thing you pick out, it may be the case that he is unable to do so – a simple (epistemic) possibility gleaned from the fact that he is not able to do everything.

> [@](#):
>
> I say, hell no. I declare that any omnipotent god has at least as much control over our universe as the writer of a book has over its contents.

That would be overreaching, since I can very well write a book in which its protagonist crosses all of Königsberg’s bridges one after the other, but god can’t make a world in which that is the case.

> [@](#):
>
> Gods who can do with the universe as they will.

The thing is, we know that god _can’t_ do with the universe as he will.

> [@](#):
>
> This does not work if you use “possible” correctly. Specifically, this says nothing at all about about whether y is a P in _our_ possible world. Which is the issue under discussion.

No, the issue under discussion is whether it’s _possible_ for y to be a P.

> [@](#):
>
> You are _grossly_ misusing “possible”. And you cannot handwave your claim that virtually nothing is physically (and thus, logically) possible. This is not seriously disputable; the laws of physics have variables in them and everything,

Which is commonly thought to be an artefact of us not yet having a complete theory of everything. String theory, for instance, contains as far as I know only one freely adjustable parameter, the string tension. It still yields a proliferation of physically possible worlds – last time I heard, on the order of 10[sup]500[/sup] of them, which is a lot, but far from infinity, meaning that yes, almost everything _is_ physically impossible. If a selection principle can be found, which is currently the hope of those string theorist not wishing to succumb to anthropic reasoning, it might well be the case that only our own world emerges as physically possible.

> [@](#):
>
> You don’t get unstated assumptions in logic. Period.

I have stated what I assume clearly before; and besides, this whole discussion wouldn’t make sense if we weren’t operating under this assumption.

> [@](#):
>
> And you apparently also don’t understand how proofs by contradiction work, either. The existence of god is **very explicitly** what in the POE is attempting to prove/disprove,

Yes, but not what I am trying to prove/disprove.

> [@](#):
>
> To disprove the POE with a Leibniz-like argument, you **must** convince your audience to accept the premise E - or more accurately, that **all** evil we see on earth is necessary, and thus that there is _no_ unnecessary evil in our world.
> 
> Just proving \<\>E gives you nothing at all.

Do we have to go through this again? I have previously shown that accepting G -\> ~E as the problem of evil, what has to be demonstrated to disprove it is G & \<\>E. Since in both cases, if G is true, the truth value does only depend on whether or not the latter half of the statement is true, I can afford not to bother with it. Putting it more succinctly, if I prove that \<\>E under all circumstances, \<\>E is true if there is a god, as well.

> [@](#):
>
> Your belief that you don’t have to prove E to make a Leibniz-like argument is based on your errors with the word “possible”. You’re wrong.

Hell, I’ve even given you a cite to that effect some hundred posts back!

> [@](#):
>
> I know exactly what I’m saying, thanks very much. If you disagree with me about what I’m saying, that’s _your_ error.

So, you’re right, and if I disagree, I’m wrong. Got it.

> [@](#):
>
> Actually I’m not sure that “In W2, \<\>R is true” makes _sense_. That is, I think it’s _explicitly_ gibberish, of the “yellow addition” variety. The way I was taught \<\> only has meaning when referring to a set of possible worlds. I mean, what do _you_ imagine “In W2, \<\>R” means?

It has a very precise sense in that in every world accessible from W2 (‘accessible from’ here meaning that one would hold to be logically possible in W2 – which in general reduces to a world that is logically possible, period), R may or may not be the case.

> [@](#):
>
> I don’t know what you imagine the POE says, but this isn’t it. And my comment was directly relevent to the blatant error you were making, Try reading it with the text it was responding to in mind and perhaps it will make more sense to you.

You’re exactly right, _the_ POE doesn’t say this. However, _your_ POE does. That’s why I don’t think it works as a POE. We agree that _your_ POE implies G -\> ~E, yes? Then, what _the_ POE is supposed to accomplish is to yield, via modus tollens, the nonexistence of god upon the assertion of the existence of evil.

However, applying this modus tollens to your POE yields merely the non-necessary existence of god; any theist could completely agree with your argument, and yet rationally continue believing in god.

> [@](#):
>
> And as I responded to that note in this thread, the Plantinga argument is about libertarian free will and is readily countered based on the fact there’s no such thing.

Actually, Plantinga’s argument merely needs for ‘if there is libertarian free will, then god can’t create a world with no evil’ to be true in order to work, not for there to actually _be_ libertarian free will. He needs, as I do, to exhibit possible circumstances in which god and evil coexist to disprove the POE; he chooses to do so by example, exhibiting a concrete world in which this is so to get to his desired conclusion. There’s nothing wrong with that, and in principle, you can buy his argument even if you disbelieve in free will (and should if you can’t find a way to show that even in a world in which there is morally significant libertarian free will, it is possible for god to keep it evil-free). Basically, all that I do differently is that I contend myself with arguing that you can always find such an example; that for every given act intended to better this world, you can always construct a (usually ridiculous) example such that this act actually makes everything worse.

Recall the Franklin quote: “To show propositions are logically compatible, all that is needed is to exhibit a merely possible scenario in which they are all true.”

This is the reason I don’t have to show that all the evil in our world is necessary, or that god actually can’t do something or create some world, but merely need to show the possibility that this is so. This is also the reason Plantinga can content himself with showing that _if_ there is morally significant free will, then god can’t create a world without evil. For Plantinga’s argument, you don’t need to grant that there _is_ free will, you merely need to grant that if there were, there’d be evil in the world, even though god is omnimax; for my argument, you don’t need to grant that a world with less evil is logically impossible, merely that it is possible that there is no such world.

This is really the crucial point that you miss whenever you assert I have to show E, or that I have to show that god can’t create a world with less evil than this one, or anything like that. I’m sorry to put it so bluntly, but as long as you do that, you just show that you don’t get what the argument is about. I really don’t think there’s any point in continuing as long as you fail to realize this.

---

<div class="post-metadata">

**Author:** ![begbert2](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/begbert2/32/3357_2.png) [@begbert2](https://boards.straightdope.com/u/begbert2)\
**Post date:** [October 7, 2010, 7:06pm UTC](https://boards.straightdope.com/t/the-best-of-all-possible-worlds/553190/266 "2010-10-07T19:06:58Z")

</div>

> [@Half\_Man\_Half\_Wit](#):
>
> Ah, I think I see what you mean now. What you call the English meaning, I would call the epistemic meaning of possibility; your logical meaning I know as the metaphysical meaning.

Every time you use the symbols \<\> and for “epistemic” possibility, it breaks my brain. I’ve never seen the symbols being used like that, I’ve never _heard_ of the symbols being used like that, and of course it has nightmarish effects on the readability of the argument to overload the symbols like that.

Especially since it’s completely redundant, since epistemic possibility is already encoded into symbolic logic of every type from the simplest to the most complex. (Epistemic )p is written ‘p’. (Epistemic )~p is written ‘~p’. And (epistemic \<\>)p and (epistemic \<\>)~p are designated by the absence of either ‘p’ or ‘~p’, or alternatively the written statement ‘p or ~p’.

Since overloading \<\> and , and the words “possible” and “necessary”, is a horrible horrible evil nasty monstrous thing to do to a poor innocent argument and its poor innocent readers, I don’t suppose you could be troubled to eschew the evil and just stick to the old-fashioned, unambigous, ‘p’/‘p or ~p’/‘~p’ way of describing epistemic possibility?

Actually, while I’m asking favors, I don’t suppose we could write this entire argument in predicate logic instead, using the variables as possible worlds? As best I can tell you can completely model modal logic that way, with A(for all) standing in for and E(there exists) for \<\>. So, the POE might be written:

G(x) = there is an omnimax god in possible world x  
E(x) = there is evil in possible world x

1: Ax.(G(x) -\> ~E(x)) _-from omnimaxness_  
2: E(earth) _-from observation_  
3: G(earth) -\> ~E(earth) _-from 1_  
4: ~G(earth) _-from 2,3_

The Leibniz response, of course, is to reject Premise 1, and instead rewrite the whole POE argument as:

G(x) = there is an omnimax god in possible world x  
QE(x) = the quantity of evil in possible world x, as a comparable orderable quantity  
NE = the minimal necessary quantity of evil

1: Ex.Ay.(QE(x) \<= QE(y)) -from QE being orderable  
2: Ay.(QE(min) \<= QE(y)) -from 1, with min for x  
3: Ax.(G(x) -\> (QE(x) = QE(min))) _-from omnimaxness_  
4: QE(earth) \> 0 _-from observation_  
5: QE(earth) \> QE(min) _-from observation_  
6: G(earth) -\> (QE(earth) = QE(min))) _-from 3_  
7: ~(QE(earth) = QE(min)) _-from 5_  
8: ~G(earth) _-from 6,7_

…and then Leibniz rejects 5, while accepting 4 and instead asserting 9: QE(earth) = QE(min), which of course means that 10: QE(min) \> 0 _-from 10, 4_. The counter of course is to criticize his arguments for rejecting 5 - a counter which I find compelling, because mysterious ways are bullshit.

Of course, I don’t know that it’s possible to write your arguments in formal predicate logic - but then, I don’t know that it’s possible to write them in formal _modal_ logic either (which for the love of cheese I hope disallows conflating metaphysical and epistemic possibility). After all, you have not yet done so.

> [@Half\_Man\_Half\_Wit](#):
>
> Leibniz’ argument establishes the metaphysical possibility of god and evil coexisting; if you prove the existence of a better world, you disprove the epistemic possibility of us living in a world set in which this metaphysical possibility is actual. Doesn’t mean that such a world set is no longer metaphysically possible any more than the observation that it’s raining means that it is metaphysically impossible that it doesn’t rain; it is only epistemically impossible.

Leibniz’ argument _attempts_ to establish the metaphysical possibility of god and evil coexisting, by baldly asserting it to be true. Or more specifically, by baldly asserting that QE(min) is nonzero, and then further baldly asserting that QE(earth) is QE(min), with an argument behind it that is pretty much only convincing to the choir.

And if one proves the mere _unactualized_ metaphysical possiblity of a min with less evil than earth, then that disproves Leibniz and disproves God.

> [@Half\_Man\_Half\_Wit](#):
>
> Even this can be gotten around by the hypothesis that relativity is logically necessary, which is a (epistemic) possibility.

But disproves omniscient Gods. Adding Ax.~G(x) (or ~G, if you prefer) to the set of knowns does the theist perspective no favors.

> [@Half\_Man\_Half\_Wit](#):
>
> Well, yes, that’s how current cosmology works – the universe wasn’t created in time, but _with_ time.

That’s what I means, sorry for the unclarity.

> [@Half\_Man\_Half\_Wit](#):
>
> That, of course, doesn’t follow.
> 
> They’d still be inconsistent with relativity, though.

It does follow, from what I meant, and if world-states are inconsistent with relativity then so is the existence of the universe. If the universe could have started with the moment of the big bang, then it could have started with the moment of last thursday, and have ended with the same moment. By “moment” here, I am definining “moment” to be whatever you think happened at the big bang. So if you get a big bang, I get world-states.

> [@Half\_Man\_Half\_Wit](#):
>
> That’s the problem! In a world state, such as, for example, my spatial hypersurface of simultaneity, there _is_ no Socrates – that word is a non-denoting name, as much as ‘the present king of France’. So there is _no_ definite truth value to the statement ‘Socrates is dead’, other than on those world slices in which it is wrong. And yet, typically, every possible world should be able to assign a truth value to this statement.
> 
> \<snip\>
> 
> It’s just that, in your modal logic, the vast majority of statements are undecidable – like ‘Socrates is dead’, for every world state in which it isn’t false.

It’s pretty easy to define “Socrates” in such a way that it is a denoting name. Same with “dead”. Even with both relativity _and_ world-states (which, as noted, must be compatible if we exist).

It’s always easy to define your terms so nothing makes sense. But in the POE, they’re defined so they do make sense. You can’t counter the POE with an argument that doesn’t address it on its own terms.

> [@Half\_Man\_Half\_Wit](#):
>
> I’m not sure why I keep trying to explain this, or why you seem to have such a hard time with it – I am not saying that he is unable to do anything at all. I’m saying that for every thing you pick out, it may be the case that he is unable to do so – a simple (epistemic) possibility gleaned from the fact that he is not able to do everything.

Except that as I believe we’ve agreed on, an epistemic possibility can be instantly eliminated by proof to the contrary. Which in this case means that any argument that shows that something is logically possible beats it.

Your argument here is basically of the form “Socrates is a man. Plato is a man. Therefore, there is an epistemic possibility that Socrates is Plato”. Which is true, as far as it goes - but it doesn’t actually go _anywhere_. It is not in any way a disproof of “Socrates is _not_ Plato”, and in fact is swept away instantly by the entry of that fact. Succinctly, an epistemic possibility is _not_ a counterargument to anything, by definition!

So, when I say “the definition of omnipotence says god can cure cancer”, and you say “but it’s epistemic possibility that he can’t”, that’s why your statement doesn’t make a dent. I have a _definition_ here, and it trumps your epistemic possibility without even slowing down. If you want to counter my definition, you have to bring something has argument _merit_ - which epistemic possibility by definition lacks.

> [@Half\_Man\_Half\_Wit](#):
>
> That would be overreaching, since I can very well write a book in which its protagonist crosses all of Königsberg’s bridges one after the other, but god can’t make a world in which that is the case.

It’s impossible to write a book in which its protagonist crosses all of Königsberg’s bridges one after the other _by following the rules_. If the god cheats, by teleporting him around or reducing the ‘crossing count’ of bridges or just declaring victory without going through the steps, those outcomes aren’t defined to be logically impossible by the Königsberg scenario.

I stand by my definition of omnipotence. But, I also invite you to provide your own! Please give a complete list of the kinds of things God can do, without elimitating the epistemic possibility that he is unable to improve the earth. I’ll let you know if that definition sounds like omnipotence to me.

> [@Half\_Man\_Half\_Wit](#):
>
> The thing is, we know that god _can’t_ do with the universe as he will.

He can get close enough to make the POE ironclad, though.

> [@Half\_Man\_Half\_Wit](#):
>
> No, the issue under discussion is whether it’s _possible_ for y to be a P.

_Epistemically_ possible? Epistemically speaking, it’s possible that I’m a talking dog. Nobody gives a shit about epistemic possibility.

> [@Half\_Man\_Half\_Wit](#):
>
> Which is commonly thought to be an artefact of us not yet having a complete theory of everything. String theory, for instance, contains as far as I know only one freely adjustable parameter, the string tension. It still yields a proliferation of physically possible worlds – last time I heard, on the order of 10[sup]500[/sup] of them, which is a lot, but far from infinity, meaning that yes, almost everything _is_ physically impossible. If a selection principle can be found, which is currently the hope of those string theorist not wishing to succumb to anthropic reasoning, it might well be the case that only our own world emerges as physically possible.

Then what means “omnipotence”?

> [@Half\_Man\_Half\_Wit](#):
>
> I have stated what I assume clearly before; and besides, this whole discussion wouldn’t make sense if we weren’t operating under this assumption.

Of course you don’t have to assume the existence of god to write the POE; neither of my predicate formulations do. And I don’t give a shit about what you stated in some rhetorical way some while ago: when you write something in _sybolic logic_, you do **not** get unstated assumptions. Ever. Period. Full stop.

This horse is dead. Stop beating it. You will never convince me that you don’t have to state what the hell you’re talking about in formal logic.

> [@Half\_Man\_Half\_Wit](#):
>
> Yes, but not what I am trying to prove/disprove.

Trying to prove that you can’t disprove god counts too.

This horse is also dead. If you’re not talking about whether god can be disproved or not, then you are not talking about the POE, and that is that.

> [@Half\_Man\_Half\_Wit](#):
>
> Do we have to go through this again? I have previously shown that accepting G -\> ~E as the problem of evil, what has to be demonstrated to disprove it is G & \<\>E. Since in both cases, if G is true, the truth value does only depend on whether or not the latter half of the statement is true, I can afford not to bother with it. Putting it more succinctly, if I prove that \<\>E under all circumstances, \<\>E is true if there is a god, as well.

Who is saying that G -\> ~E is the POE? I’ve never seen a formulation where that is the case. In all the _valid_ formulations of it I’ve seen, you never mix symbols with and without modal tags on them. Heck, I’m not even sure what the resulting sentences are supposed to _mean_. What _does_ G -\> ~E mean? The existence of god implies that there no evil in any possible world? The existence of God where? In one possible world? In all possible worlds? Somewhere in the middle?

Seriously - modal logic sybology is bullshit anyway; I’ve literally never seen it used except in making fallacious religious arguments. But if you’re going to use it, at least be consistent about it.

> [@Half\_Man\_Half\_Wit](#):
>
> So, you’re right, and if I disagree, I’m wrong. Got it.

About the meaning of my on words? Yes, that’s right. When I say something and you misinterpret it, then you’re wrong. Life’s tough that way.

> [@Half\_Man\_Half\_Wit](#):
>
> It has a very precise sense in that in every world accessible from W2 (‘accessible from’ here meaning that one would hold to be logically possible in W2 – which in general reduces to a world that is logically possible, period), R may or may not be the case.

Ah, okay, that ‘accessibility’ garbage. Didn’t I dismiss that as irrelevent a few billion posts ago? God’s not limited to what accessible from here, and there’s literally no way that that limited set of worlds can possibly be relevent to the discussion. So there’s no need to further obfuscate the situation by introducing such things.

If you _actually_ want to introduce something like accessible worlds, use something like predicate logic where they can be defined and tracked unambiguously. Unless, of course, obfusaction is the point.

> [@Half\_Man\_Half\_Wit](#):
>
> You’re exactly right, _the_ POE doesn’t say this. However, _your_ POE does. That’s why I don’t think it works as a POE. We agree that _your_ POE implies G -\> ~E, yes? Then, what _the_ POE is supposed to accomplish is to yield, via modus tollens, the nonexistence of god upon the assertion of the existence of evil.
> 
> However, applying this modus tollens to your POE yields merely the non-necessary existence of god; any theist could completely agree with your argument, and yet rationally continue believing in god.

Didn’t we already establish that you are mistaken about what I was talking about?

Look, let’s back up to before all this. Why don’t _you_ tell me what version of the POE you want to talk about. Its precise arguments, its precise formulation. Because there are so many different variants of the thing that we’re very likely talking past one another. Heck, I provided two more of them at the top of this very post!

You tell me what version of the POE you’re disproving, and how you’re disproving it, and I’ll tell you whether your argument works or not. Let’s keep it self-contained and simple, so no other variant POEs and counterarguments slip in. Okay?

> [@](#):
>
> And as I responded to that note in this thread, the Plantinga argument is about libertarian free will and is readily countered based on the fact there’s no such thing.

Actually, Plantinga’s argument merely needs for ‘if there is libertarian free will, then god can’t create a world with no evil’ to be true in order to work, not for there to actually _be_ libertarian free will. He needs, as I do, to exhibit possible circumstances in which god and evil coexist to disprove the POE; he chooses to do so by example, exhibiting a concrete world in which this is so to get to his desired conclusion. There’s nothing wrong with that, and in principle, you can buy his argument even if you disbelieve in free will (and should if you can’t find a way to show that even in a world in which there is morally significant libertarian free will, it is possible for god to keep it evil-free). Basically, all that I do differently is that I contend myself with arguing that you can always find such an example; that for every given act intended to better this world, you can always construct a (usually ridiculous) example such that this act actually makes everything worse.

Recall the Franklin quote: “To show propositions are logically compatible, all that is needed is to exhibit a merely possible scenario in which they are all true.”

This is the reason I don’t have to show that all the evil in our world is necessary, or that god actually can’t do something or create some world, but merely need to show the possibility that this is so. This is also the reason Plantinga can content himself with showing that _if_ there is morally significant free will, then god can’t create a world without evil. For Plantinga’s argument, you don’t need to grant that there _is_ free will, you merely need to grant that if there were, there’d be evil in the world, even though god is omnimax; for my argument, you don’t need to grant that a world with less evil is logically impossible, merely that it is possible that there is no such world.

This is really the crucial point that you miss whenever you assert I have to show E, or that I have to show that god can’t create a world with less evil than this one, or anything like that. I’m sorry to put it so bluntly, but as long as you do that, you just show that you don’t get what the argument is about. I really don’t think there’s any point in continuing as long as you fail to realize this.  
[/QUOTE]  
Epistimic possibility is argumentively worthless - to the point of not even being worth _mentioning_. Which is why your ‘crucial point’ is no point at all.

And no, you can’t defend Plantinga that way. “If there were an elephant on your head, then you’d be dead. There isn’t one, but it’s still possible you’re dead anyway, despite evidence to the contrary,” is not good argument.

---

<div class="post-metadata">

**Author:** ![PBear42](https://avatars.discourse-cdn.com/v4/letter/p/ecccb3/32.png) [@PBear42](https://boards.straightdope.com/u/PBear42)\
**Post date:** [October 8, 2010, 6:15am UTC](https://boards.straightdope.com/t/the-best-of-all-possible-worlds/553190/267 "2010-10-08T06:15:23Z")

</div>

Checking in, I’m following the thread as an edification exercise in formal logic. I thought Leibniz’s and the **HMHW’s** argument is an epistemological proposition, that we can’t KNOW (with certainty) that this isn’t the best possible world an omnimax god can create. As a materialist, I can grant this under the ground rules of formal logic and not care, just as I don’t care about brain-in-a-vat. (Silly me, I think induction means something.) In his latest post, **HMHW** asserts “Leibniz’ argument establishes the metaphysical possibility of god and evil coexisting.” Well, if true, that would be interesting, but I haven’t seen it (though, admittedly, I’m working only from secondary sources). What’s the metaphysical argument?

---

<div class="post-metadata">

**Author:** ![PBear42](https://avatars.discourse-cdn.com/v4/letter/p/ecccb3/32.png) [@PBear42](https://boards.straightdope.com/u/PBear42)\
**Post date:** [October 8, 2010, 7:15am UTC](https://boards.straightdope.com/t/the-best-of-all-possible-worlds/553190/268 "2010-10-08T07:15:10Z")

</div>

BTW, **begbert2** , since I’m posting, I’'d like to point out again that just because you think free will is dead letter doesn’t make it so. For example, the [IEP article on free will](http://www.iep.utm.edu/freewill/) summarizes the current state of the debate as follows:

> [@](#):
>
> Most philosophers agree that whether or not determinism is true is a contingent matter; that is, determinism is neither necessarily true nor necessarily false. If this is so, then whether or not determinism is true becomes an empirical matter, to be discovered by investigating the way the world is, not through philosophical argumentation.

Which is to say that Plantinga’s defence passes the possible test. Whether it’s true is another matter. And whether it meets the evidential POE is yet another (here, I think he fails massively, and that as someone who believes in meaningful volition).

---

<div class="post-metadata">

**Author:** ![begbert2](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/begbert2/32/3357_2.png) [@begbert2](https://boards.straightdope.com/u/begbert2)\
**Post date:** [October 8, 2010, 3:48pm UTC](https://boards.straightdope.com/t/the-best-of-all-possible-worlds/553190/269 "2010-10-08T15:48:34Z")

</div>

> [@PBear42](#):
>
> BTW, **begbert2** , since I’m posting, I’'d like to point out again that just because you think free will is dead letter doesn’t make it so. For example, the [IEP article on free will](http://www.iep.utm.edu/freewill/) summarizes the current state of the debate as follows:
> 
> > [@](#):
> >
> > Most philosophers agree that whether or not determinism is true is a contingent matter; that is, determinism is neither necessarily true nor necessarily false. If this is so, then whether or not determinism is true becomes an empirical matter, to be discovered by investigating the way the world is, not through philosophical argumentation.
> 
> Which is to say that Plantinga’s defence passes the possible test. Whether it’s true is another matter. And whether it meets the evidential POE is yet another (here, I think he fails massively, and that as someone who believes in meaningful volition).

Well, my argument against free will isn’t dependent on deterministm vs. nondeterminism per se; it’s based on a computational understanding that cognition is necessarily carried out by processes that can be distilled into the two categories “determined” and “random” with no “magical sentient atomic free will particles” left over. And observational analysis that people are mostly capable of walking in a straight line proves that what randomity there is, if any, is largely corrected for by the determined processes anyway. (I’m a computer scientist; calculative processes are my forte.) I am not under the impression that the philosophers entertaining libertarian free will are accounting for this line of argument.

Regardless, Plantinga’s argument, as best I understand it, relies on God having a _different_ kind of free will than ours, an omnibenevolent one - which leaves one asking how he was able to omnibenevolently give us broken crappy non-omnibenevolent free will, when something that would create a better result is available. If Plantinga answered this, I guess I missed it.

---

<div class="post-metadata">

**Author:** ![PBear42](https://avatars.discourse-cdn.com/v4/letter/p/ecccb3/32.png) [@PBear42](https://boards.straightdope.com/u/PBear42)\
**Post date:** [October 9, 2010, 5:20am UTC](https://boards.straightdope.com/t/the-best-of-all-possible-worlds/553190/270 "2010-10-09T05:20:26Z")

</div>

Actually, Plantinga’s argument doesn’t have anything to do with God’s free will. That’s just a point that has come up in the logical POE debate and I mentioned it (many days ago) as it resembled an argument **Voyager** had made. Rather, the fundamental thesis of Plantinga’s defense, as quoted in the [IEP article](http://www.iep.utm.edu/evil-log/), is this:

> [@](#):
>
> God’s creation of persons with morally significant free will is something of tremendous value. God could not eliminate much of the evil and suffering in this world without thereby eliminating the greater good of having created persons with free will with whom he could have relationships and who are able to love one another and do good deeds.

Although framed in terms of LFW, there’s no obvious reason this is necessary. I’m pretty sure my thesis of soft or psychological compatabilism would be sufficient. As would most forms of traditional compatabilism of which I’m aware. For that matter, so would your process thesis, if you allow for moral responsibility.

As I said, I think Plantinga’s defense runs into real trouble in the context of the evidential POE, including for the reason mentioned in your next-to-last sentence. A bigger problem, IMHO, are the aspects of human nature that make so many of us such crappy people. For purposes of a logical defense, though, my understanding is that none of this matters. See § 7 of the IEP article.

---

<div class="post-metadata">

**Author:** ![Half\_Man\_Half\_Wit](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/half_man_half_wit/32/21766_2.png) [@Half\_Man\_Half\_Wit](https://boards.straightdope.com/u/Half_Man_Half_Wit)\
**Post date:** [October 15, 2010, 8:54am UTC](https://boards.straightdope.com/t/the-best-of-all-possible-worlds/553190/271 "2010-10-15T08:54:59Z")

</div>

> [@begbert2](#):
>
> Every time you use the symbols \<\> and for “epistemic” possibility, it breaks my brain. I’ve never seen the symbols being used like that, I’ve never _heard_ of the symbols being used like that, and of course it has nightmarish effects on the readability of the argument to overload the symbols like that.

Well, epistemic logic is one of the more common subspecies of modal logic – it even made the [wiki page](http://en.wikipedia.org/wiki/Modal_logic#Epistemic_logic).

> [@](#):
>
> Especially since it’s completely redundant, since epistemic possibility is already encoded into symbolic logic of every type from the simplest to the most complex. (Epistemic )p is written ‘p’. (Epistemic )~p is written ‘~p’. And (epistemic \<\>)p and (epistemic \<\>)~p are designated by the absence of either ‘p’ or ‘~p’, or alternatively the written statement ‘p or ~p’.

Epistemic statements are statements about the knowledge of an agent (or a group of agents), so it can be the case that \<\>p, even if ~p.

> [@](#):
>
> Since overloading \<\> and , and the words “possible” and “necessary”, is a horrible horrible evil nasty monstrous thing to do to a poor innocent argument and its poor innocent readers, I don’t suppose you could be troubled to eschew the evil and just stick to the old-fashioned, unambigous, ‘p’/‘p or ~p’/‘~p’ way of describing epistemic possibility?

Actually, I’m not sure I could. It’s not easy to capture the whole of epistemic logic in predicate logic, and I’d probably bungle it up somewhere.

> [@](#):
>
> Actually, while I’m asking favors, I don’t suppose we could write this entire argument in predicate logic instead, using the variables as possible worlds? As best I can tell you can completely model modal logic that way, with A(for all) standing in for and E(there exists) for \<\>.

Well, you can recast any modal argument in predicate logic, in principle, using what’s called the [standard translation](http://en.wikipedia.org/wiki/Standard_translation); however, this requires talk about accessibility relations, which you don’t seem to like for some reason.

> [@](#):
>
> Leibniz’ argument _attempts_ to establish the metaphysical possibility of god and evil coexisting, by baldly asserting it to be true. Or more specifically, by baldly asserting that QE(min) is nonzero, and then further baldly asserting that QE(earth) is QE(min), with an argument behind it that is pretty much only convincing to the choir.

Neither is asserted; both are argued to be possible. This really is the key point.

> [@](#):
>
> It does follow, from what I meant, and if world-states are inconsistent with relativity then so is the existence of the universe. If the universe could have started with the moment of the big bang,

First, you say that the universe being created with time is what you mean; then, you say this, which is wildly contradictory. There is no good notion of the ‘moment’ of the big bang, anymore than there is a good notion of the location of the big bang.

> [@](#):
>
> then it could have started with the moment of last thursday, and have ended with the same moment.

Regardless of everything else, this still doesn’t follow. You’ll have to get here from there, and there might, for all we know, be only one way to do so. Claiming that the existence of one way for the universe to evolve to its present state implies the existence of another is like claiming that the sun rose yesterday implies the sun rises tomorrow; it’s just an inductive fallacy.

> [@](#):
>
> By “moment” here, I am definining “moment” to be whatever you think happened at the big bang. So if you get a big bang, I get world-states.

What I think happened at the big bang really doesn’t have anything to do with how one could define moment sensibly to obtain a state of the world. Really, you fail to appreciate just how soundly relativity abolishes naive Newtonian notions of time.

> [@](#):
>
> It’s pretty easy to define “Socrates” in such a way that it is a denoting name.

But not in such a way as to refer to the individual you and I mean when talking about Socrates – as that doesn’t exist in your world state.

> [@](#):
>
> Except that as I believe we’ve agreed on, an epistemic possibility can be instantly eliminated by proof to the contrary. Which in this case means that any argument that shows that something is logically possible beats it.

What we learn if you show that my car is indeed black is that we live in a world-set in which it is possible for my car to be black. What we don’t learn is whether it is impossible for my car to be white; i.e. of the three possible world sets in which 1) my car is black in all worlds, 2) my car is white in some worlds and black in others (ignoring other colourings for the moment), 3) my car is white in all worlds, we can eliminate 3). Leibniz’ counter to the POE asserts that we may live in 2); the POE asserts we live in 1). So pointing out my car’s blackness doesn’t decide that question (though it does, of course, decide the question of what colour my car actually _is_; same as showing that it is possible for the world to contain less evil decides whether or not god actually is omnimax – he isn’t, in that case).

> [@](#):
>
> Your argument here is basically of the form “Socrates is a man. Plato is a man. Therefore, there is an epistemic possibility that Socrates is Plato”. Which is true, as far as it goes - but it doesn’t actually go _anywhere_. It is not in any way a disproof of “Socrates is _not_ Plato”, and in fact is swept away instantly by the entry of that fact. Succinctly, an epistemic possibility is _not_ a counterargument to anything, by definition!

Well, it is a counterargument to any epistemic claim, of course. One could claim that from ‘Socrates is a man’ and ‘Plato is a man’, one could conclude that ‘Plato is not Socrates’, but using epistemic logic, it is rather easy to show that this doesn’t follow – with just these two bits of information, you can’t conclude that Socrates isn’t Plato. As you say, you have to bring in extra info to do so; however, even after you’ve done that, the first argument still remains fallacious. So, as far as the POE entails a claim about knowing which worlds are possible, it is only fair to point out that our knowledge is actually not good enough (and can’t be) to make that claim.

> [@](#):
>
> So, when I say “the definition of omnipotence says god can cure cancer”, and you say “but it’s epistemic possibility that he can’t”, that’s why your statement doesn’t make a dent. I have a _definition_ here,

No. What you have is an overreaching assertion of knowledge.

> [@](#):
>
> It’s impossible to write a book in which its protagonist crosses all of Königsberg’s bridges one after the other _by following the rules_. If the god cheats, by teleporting him around or reducing the ‘crossing count’ of bridges or just declaring victory without going through the steps, those outcomes aren’t defined to be logically impossible by the Königsberg scenario.

They are also not shown to be logically possible by your saying so.

> [@](#):
>
> I stand by my definition of omnipotence. But, I also invite you to provide your own! Please give a complete list of the kinds of things God can do,

That’s the thing – I can’t. And neither can you. My definition of omnipotence is the ability to do anything that can be done. I just don’t know what exactly that is, and, despite your continuing assertions to the contrary, neither do you.

> [@](#):
>
> _Epistemically_ possible? Epistemically speaking, it’s possible that I’m a talking dog.

It’s not. I know, for example, that dogs lack the vocal apparatus as well as the cortical sophistication to produce speech. So I know there are no talking dog. So I know you are not one. See? Epistemic argument, nevertheless it established something about the real world. There’s great use to epistemic logic, and your denial of that, frankly, seems a bit odd.

> [@](#):
>
> This horse is dead. Stop beating it. You will never convince me that you don’t have to state what the hell you’re talking about in formal logic.

I’ve stated it very clearly; and for a cite that you don’t have to carry around assumptions as long as they don’t change, here’s the wiki on [modus tollens](http://en.wikipedia.org/wiki/Modus_tollens); see the formulation including the gammas. Now that’s of course not what I’ve been doing, but it shows that you _do_ get unstated assumptions, at least in some cases.

> [@](#):
>
> Trying to prove that you can’t disprove god counts too.

Well, I’m not trying to do that, either. I’m just trying to show that the POE fails.

> [@](#):
>
> This horse is also dead. If you’re not talking about whether god can be disproved or not, then you are not talking about the POE, and that is that.

Come on, of course I can attack an argument without needing to talk about its conclusion! I can show the premises faulty, or find a formal flaw, neither of which necessitates any talk about the content of the argument.

> [@](#):
>
> Who is saying that G -\> ~E is the POE? I’ve never seen a formulation where that is the case. In all the _valid_ formulations of it I’ve seen, you never mix symbols with and without modal tags on them.

Then you probably just haven’t seen very many modal arguments. Here’s [one](http://en.wikipedia.org/wiki/G%C3%B6del%27s_ontological_proof) with quite liberal mixing; and hell, even the axioms of modal logic contain terms with and without operators! So I’m really not sure where you get this ‘non-mixing’ from, and frankly, it casts some doubts on your claimed mastery of modal logic.

> [@](#):
>
> Heck, I’m not even sure what the resulting sentences are supposed to _mean_. What _does_ G -\> ~E mean? The existence of god implies that there no evil in any possible world? The existence of God where? In one possible world? In all possible worlds? Somewhere in the middle?

Just the existence of god, nothing else. There’s really no controversy there.

> [@](#):
>
> Ah, okay, that ‘accessibility’ garbage. Didn’t I dismiss that as irrelevent a few billion posts ago?

Well, it is an elementary concept in the valuation of modal logic, so your dismissal doesn’t really mean terribly much.

> [@](#):
>
> Didn’t we already establish that you are mistaken about what I was talking about?

Well, you did claim that I was, without backing it up in any way. Basically, you proposed a version of the POE in modal symbology, which I don’t think makes the cut, and made an argument to that effect, showing that it implies a situation in which god possibly exists, and there is evil in this world. You then said that that was not what you meant, and best as I can gather, you seem to just mean that you meant for your symbology to capture the actual POE; but that doesn’t change the fact that it doesn’t.

> [@](#):
>
> Look, let’s back up to before all this. Why don’t _you_ tell me what version of the POE you want to talk about. Its precise arguments, its precise formulation.

Damn, again? Oh well: The argument can be formalized as modus tollens from G -\> ~E; so that, upon asserting E, ~G follows (and not, as in your version, ~G).

> [@](#):
>
> You tell me what version of the POE you’re disproving, and how you’re disproving it, and I’ll tell you whether your argument works or not. Let’s keep it self-contained and simple, so no other variant POEs and counterarguments slip in. Okay?

Disproof works by showing ~(G -\> ~E) \<-\> G & \<\>E.

Basically, we may live in one of three world-sets:

1. in which, whenever there is a god, there is no evil;
2. in which there may or may not be evil when there is a god; and
3. in which there is always evil when there is a god.

The POE asserts that we live in 1); alternatively (but equivalently), it asserts that the premises a) god is omnipotent, b) god is omniscient, c) god is omnibenevolent, d) there exists evil are incompatible, since a contradiction can be derived from them. (From a), b), c) it follows that there is no evil, from d) it follows, trivially, that there is.)

This is a claim that is both metaphysical and epistemic: it asserts that we both live in 1), and know that we do (by deriving the contradiction from a, b, c, and d). The Leibniz argument now intends to show that (for all we know), we may live in either of the world sets.

In my version, it does so by noting that there are some things that god can’t accomplish (the logically impossible things), and that a priori, there is no way for us to know which things these are, by, for instance, noting that for each way that can be thought of to improve this world, one can think of a way in which this improvement might make matters worse, or be flat impossible. Plantinga’s version gives a concrete example of such an allegedly impossible thing, the existence of morally significant free will without the possibility of evil things happening (out of interest, do you agree that granting Plantinga free will, his argument works?).

The key thing to realise is that Leibniz’ account in no way settles the question of whether we live in 1), 2), or 3), and doesn’t aim to – it merely argues that it is possible for us to live in 2) (or even 3), and hence, for the metaphysical claim of the POE that we live in 1) to be false. Showing ‘we can’t know whether we live in 1), 2), or 3) a priori’ amounts to showing the POE false, as it asserts that we can know that we live in 1). Of course, it is easy to empirically establish that we don’t live in 1) or 3), but that doesn’t tell us anything about whether or not we live in 2); so as long as it’s (epistemically) possible for us to live in 2), the POE doesn’t go through. This epistemic possibility is established by noting the possibility of any given thing to be impossible to accomplish even for an omnipotent being.

> [@](#):
>
> And no, you can’t defend Plantinga that way. “If there were an elephant on your head, then you’d be dead. There isn’t one, but it’s still possible you’re dead anyway, despite evidence to the contrary,” is not good argument.

And doesn’t have anything to do with the argument actually being made.

[Previous page](https://boards.straightdope.com/t/the-best-of-all-possible-worlds/553190.md?page=13)
