Evil and Compatibilism
February 8, 2015 — 11:33

Author: Michael Almeida  Category: Concept of God Existence of God Free Will General Problem of Evil Uncategorized  Tags: , , , ,   Comments: 17

There is widespread belief that compatibilism + theism cannot offer a credible solution to the logical problem of evil. Why does anyone believe that? I think they’re reasoning this way: if compatibilism is true, then, necessarily, God can actualize a morally perfect world. That’s of course true, and it entails that the free will defense fails. But then they reason, if, necessarily, God can actualize a morally perfect world, then, necessarily, God does actualize a morally perfect world. It is then observed that, obviously, there is evil. So, compatibilism + theism is incoherent; it cannot solve the logical problem.

But the inference from, necessarily, God can actualize a morally perfect world to, necessarily, God does actualize a morally perfect world is invalid on the assumption of compatibilism (and, for that matter, libertarianism): that is, it invalid on both Lewisian compatibilism and on any stronger compatibilism. In fact, theism + compatibilism, of either sort, is inconsistent with (M), which is the strong modal conclusion of the logical problem of evil. So, here are the choices: Either God cannot actualize a morally and naturally perfect world or the principle in (M) is false. But, surely, God can actualize a morally and naturally perfect world, I think is the right rejoinder.

M. Necessarily, God does not coexist with a single instance of moral or natural evil.

Here is the argument for the inconsistency between theism + compatibilism and (M). Let w be a world in which God arranges history H and the laws of nature L so that it is causally determined that S does A at t. Let A be a morally significant action such as keeping a promise. It is, we are assuming, all-in right to keep the promise and all-in wrong not to do so. Here is Lewisian compatibilism.

LC. S freely performed A at t only if (i) had S performed ~A at t, then it would have been the case that L is not the conjunction of laws of nature and (ii) S was able to perform ~A at t.

We’ve assumed that God creates S in world w and determines S to do the morally right action A at t. But it follows from (LC) that S is free in w, in the Lewisian sense of being free, only if there is another world w’ in which S performs a moral evil ~A at t. In w’ S brings about the morally evil action of breaking a promise that S had an all-in obligation to keep. Here’s our conclusion.

C. If God creates S in w and determines S to go right, then if S is Lewis-free in w, then there is some possible world w’ in which (a) God exists and (b) there is moral evil.

(C) is of course inconsistent with (M). So, if God is permitted to create morally perfect worlds in which everyone is causally determined to go right, and everyone is LC-free, then there must be a world in which God and moral evil coexist. But then so much for (M). If God can create a morally and naturally perfect world in which every agent is Lewis-free, then (M) is false.

But what if we assume that agents are not Lewis free. Suppose we assume stronger compatibilism, (SC).

SC. S freely performed A at t only if (i) had S performed ~A at t, then it would have been the case that L is not the conjunction of laws of nature (ii’) S was unable perform ~A at t.

Unlike (LC), (SC) does not assume that S was able to perform ~A at t. But (SC) holds that S is free anyway, despite being unable to act otherwise at t. But even on (SC) we can reach our conclusion in (C)! The conclusion in (C) follows from condition (i) alone of both (LC) and (SC). It does not matter whether we believe that, in w, S was able to perform ~A at t or not. All we need is that there is some possible world w’ in which S does perform ~A at t, and condition (i) ensures that there is such a world.

So, again, (C) is inconsistent with (M). If God is permitted to create morally perfect worlds in which everyone is causally determined to go right, and everyone is SC-free, then there must be a world in which God and moral evil coexist. So, we reach the conclusion that either God cannot actualize a morally and naturally perfect world–a world with free moral agents of some sort, compatibilist or otherwise– or the principle in (M) is false.

Comments:
  • Drew

    I think that M is a straw man argument here. It should be revised to say that Necessarily, God cannot coexist with both a deterministic world and the existence of evil. I believe this.

    Your argument shows only that under theism and Lewisian compatibilism, there can be no free moral acts. The same holds for stronger compatibilism.

    February 8, 2015 — 21:35
  • Michael Almeida

    M is the consequence of the logical problem of evil, it is not something I proposed. Whether or not there are indeterministic worlds, M follows from the premises of the logical problem. If you want to reject M, then you’ll need come up with some reason to reject some premise in that argument, since it is in fact valid.

    February 9, 2015 — 7:37
  • Anonymous

    I’m not sure I’m following your argument in the last 3 paragraphs. Why does (i) guarantee that there is a possible world where God coexists with moral evil? (i) is just an if-then statement about worlds where S performs ~A, not an assertion about the existence of such worlds. (i) could be vacuously true if there are no such worlds–L is not the conjunction of the laws of nature in any world w’ where S performs ~A, which is perfectly consistent with the assertion that there is no such world w’.

    February 12, 2015 — 16:33
  • Michael Almeida

    We have assumed that (i) is true, since S is free. I took it to be obvious that the closest worlds to w in which S does ~A at t are not impossible worlds. We have just three possibilities: the closest world is impossible, or the closest world is one where L does not conjoin the laws or the closest world is one where the history H prior to t is different. Of those three, it is evident that the closest is the second disjunct.

    February 12, 2015 — 16:47
  • Anonymous

    “We have just three possibilities: the closest world is impossible, or the closest world is one where L does not conjoin the laws or the closest world is one where the history H prior to t is different. Of those three, it is evident that the closest is the second disjunct.”

    Why is it evident that the closest world is not impossible? (You seem to be assuming this rather than arguing for it.) If God necessarily creates a morally perfect world, then necessarily, God does not create any worlds w’ where S performs ~A at t, regardless of how close they are to the morally perfect world w where S performs A at t. Thus all such worlds w’ are not possible worlds.

    February 13, 2015 — 16:54
  • Michael Almeida

    Right, the question is one of similarity. What is more similar to w, a world in which the laws are slightly different, or a world in which impossible things are true? Your suggestion seems to be that an impossible world would be more similar to w than a world in which the laws are slightly different. Impossible worlds are wildly dissimilar: everything is true in them, including contradictions, nothing is false. I claim that impossible worlds are not, therefore, as similar to w as a world in which the laws are slightly different. The line I’m taking is due to David Lewis, ‘Are We Free to Break Laws?’. That’s pretty much all I have to say about clause (ii).

    On a second point, God cannot necessarily actualize a morally perfect world. That’s metaphysically impossible. His attempt to do so would preclude the existence of free agents. But there must be free agents in a morally perfect world.

    February 13, 2015 — 17:02
  • Anonymous

    “What is more similar to w, a world in which the laws are slightly different, or a world in which impossible things are true? Your suggestion seems to be that an impossible world would be more similar to w than a world in which the laws are slightly different.”

    Right, the suggestion is that any world where S performs ~A at t (this includes some worlds where L is the conjunction of the laws of nature and some worlds where it is not) is an impossible world because God cannot actualize morally imperfect worlds.

    “On a second point, God cannot necessarily actualize a morally perfect world. That’s metaphysically impossible. His attempt to do so would preclude the existence of free agents.”

    Doesn’t that fail to hold under Stronger Compatibilism? Under SC, S can freely perform A at t even though S could not have performed ~A at t (“couldn’t have done otherwise”). So “S necessarily performs A at t” does not contradict “S freely performs A at t.”

    February 13, 2015 — 17:17
  • Michael Almeida

    Condition (ii’) on SC says that S was unable to do otherwise, not that it is metaphysically impossible that he do otherwise. There are lots of things we are unable to do that are metaphysically possible. Otherwise, (ii’) would not be consistent with (i), which ensures that it’s metaphysically possible that S fails to do A at t.

    February 13, 2015 — 17:27
  • Anonymous

    “(i)…ensures that it’s metaphysically possible that S fails to do A at t”

    I think this is what I’m having trouble understanding. Suppose that it is metaphysically impossible that S performs ~A at t. By (i), we have that if S performs ~A at t, then L is not the conjunction of the laws of nature. By hypothesis, S (necessarily) does not perform ~A at t, but why does this contradict (i)? We still have that L is not the conjunction of the laws of nature in every world where S performs ~A at t–it just so happens that there are no such worlds.

    February 13, 2015 — 17:40
  • Michael Almeida

    If it is metaphysically impossible that S performs ~A at t, then S is not free in any sense, compatibilist or not. The claim is inconsistent with (i) for reasons I’ve already been over.

    February 14, 2015 — 10:14
  • Michael Almeida

    I’ll answer this last question. But I consider the ceaseless repetition of more or less the same questions trolling. I’ve spammed the comment.

    The impossibility of ~A at t is inconsistent with (i). (i) is non-trivially true, as I show above, since impossible worlds in which ~A is true are much farther away that ~A worlds where L does not collect the laws of nature (of w).

    Worlds in which it is metaphysically impossible to do otherwise are necessitarian worlds. These worlds are taken as paradigmatic of worlds in which agents are not free–they are Spinozistic worlds. If agents are free in necessitarian worlds, then there is no genuine difference between agents that are free and those that are not.

    February 15, 2015 — 9:08
  • If agents are free in necessitarian worlds, then there is no genuine difference between agents that are free and those that are not.

    Hi Michael. But on a wide range of compatibilist theories of free will, the differences between free and unfree actions (or agents) is not metaphysical, but psychological. What matters for freedom isn’t any sort of sourcehood or ability to do otherwise (even on a compatiblist-friendly analysis of such an ability), but the psychological mechanisms that produced the person’s actions. I don’t want to go over all of the different sorts of theories, but let’s say that you endorse a view on which freedom requires a certain sort of reasons-responsiveness, or one (as Frankfurt thinks) that requires acting from second-order desires. It seems to me that that there are genuine differences between the free and unfree agents on such theories, and the differences that these theories point out don’t disappear even if it turns out that this world is necessitarian (because of considerations of God’s nature).

    February 16, 2015 — 10:27
  • Michael Almeida

    Hi Tim,

    But on a wide range of compatibilist theories of free will, the differences between free and unfree actions (or agents) is not metaphysical, but psychological.

    I didn’t say that the differences were metaphysical. I have one commitment in the post that expresses a universal proposition on free will and it is this: if it is true that our actions are metaphysically necessitated, then we are not free, I don’t care what psychological facts hold (van Inwagen argues persuasively that having an alternative is just what it means to be free; compatibilist or not, that’s what we are all talking about when it comes to discussions of free will). So, I don’t think that making it a necessary condition on free action that it is not metaphysically necessitated is controversial. And, even if someone is willing to take the position that agents in Spinozistic worlds are actually free (Spinoza wisely didn’t), I’m happy to allow him to have his bizarre belief. I can’t argue with every fringe belief about free will. Otherwise, I say, there are two views on free will that matter to my discussion. On one view, the fact that one’s actions are causally determined does not preclude alternative possibilities of action, on the other it does. But it does not matter to me, so long as on both views the agents are (purportedly) free.

    So, I guess I’d challenge you to present one credible view on free will according to which (i) one’s free action depends on intrinsic psychological facts alone (no sources, no alternatives, no non-manipulation, etc.) and (ii) the view does not implicitly assume that one’s actions are not metaphysically necessitated.

    February 16, 2015 — 10:43
  • So, I guess I’d challenge you to present one credible view on free will according to which (i) one’s free action depends on intrinsic psychological facts alone (no sources, no alternatives, no non-manipulation, etc.) and (ii) the view does not implicitly assume that one’s actions are not metaphysically necessitated.

    I think that lots (though far from all!) of contemporary compatibilist theories reject PAP and the idea that “having an alternative” is what’s central to having free will. So just as a matter of sociology of philosophy, it’s far from a fringe view to reject PAP, which you’re assuming is a shared assumption between the libertarian and compatibilist about free will. See http://plato.stanford.edu/entries/compatibilism/#5 sections 5.3-5.6–I’d argue that lots of those views fit (i) and (ii) above, although I’m open to hear arguments to the contrary. Some of the views are ahistorical (Frankfurt, Wolf), although I think a historical view is more plausible. But a historical view (IMHO) can avoid relying on an implausible of ‘sourcehood’ that’s ruled out by determinism, and can also distinguish between genuine manipulation and the supposed ‘manipulation’ cases brought up by people like Pereboom.

    February 17, 2015 — 9:57
  • Michael Almeida

    Tim,

    I did not say that only fringe compatibilists reject PAP. What I called a fringe position is one which claims that free will is compatible with necessitarianism. And that is fringy. Indeed, I could not name one compatibilist who expressly takes the position that agents in necessitarian worlds are free. So, the assumption is typically implicit. So, take a typical Frankfurt case. There is no one who finds it intuitive to conclude that, in those cases, Jones freely and responsibly goes wrong under the additional assumption that the agent’s actions are metaphysically necessitated. They have rather the intuition that Jones is not free. They have the same intuition concerning Frankfurt cases under the assumption that Black’s device, unknown to Black or Jones, was actually causing Jones to go wrong. The counterexamples to PAP (though PAP is not directly relevant to free will) all assume that the device is not causing the agent to go wrong, and they certainly assume that the agent is not metaphysically necessitated to go wrong.

    In necessitarian worlds, frankly, it is difficult to see moral ‘agents’ as performing actions. It’s more accurate to say that events are happening to them. All of their ‘actions’ are like falling.

    February 17, 2015 — 10:10
    • Hi Michael.

      Well, it’s true that no contemporary compatibilists (as far as I know) explicitly affirm that free will is compatible with metaphysical necessitarianism, although I don’t think that any explicitly deny it either—I just don’t think it’s a topic that’s much discussed in the contemporary literature. So the best we can do is to draw out what they should say about it, given their other commitments.

      So, here is my thinking: imagine that I’m committed to a reasons-responsiveness compatibilism. I believe that my action of writing this blog comment is free and responsible, whereas the actions of a person in the middle of a psychotic break, a small child, and a dog piddling on my rug are not, and I spell out what I take to be the relevant differences between these cases in terms of how I am responsive to reasons in the appropriate sort of way and they are not.

      And then somebody comes along and says “Hah—you think your actions are free; but really they’re not free at all! They’re caused by your brain states.” I’m not going to be phased by this at all, unless the person can spell out why “being caused by my brain states” entails that I’m not reasons-responsive in the way I think I am. As long as I am reasons-responsive in a way that the others aren’t, the “brain states” thesis is irrelevant.

      So why can’t I give the same sort of flat-footed response to the Spinoza advocate? I have a story to tell about the differing psychological mechanisms that produce action in my case vs. the other cases. The Spinoza thesis is a surprising bit of metaphysics—all sorts of things that I thought were metaphysically possible turn out not to be, because of weird modal properties of God and his relationship to the world. That’s interesting, but I don’t see why I can’t stick to my story about what’s going on my case vs. the others.

      Here is another way of putting it. You write,

      In necessitarian worlds, frankly, it is difficult to see moral ‘agents’ as performing actions. It’s more accurate to say that events are happening to them. All of their ‘actions’ are like falling.

      I just don’t see how this follows. Even if the actual world is the only possible world, all of the same differences between my writing this blog comment now vs. a rock falling downward are still there, exactly the same as before. I deliberate, I have beliefs and desires, I consider reasons, etc., all of which aren’t true of the rock.

      February 17, 2015 — 12:38
  • Michael Almeida

    I spell out what I take to be the relevant differences between these cases in terms of how I am responsive to reasons in the appropriate sort of way and they are not

    But you are no more reasons responsive than they are. You are no more reasons responsive than the stone is. What is true is that every action of yours is itself necessary. It has the de re modal property of necessarily occurring. Your writing the blog comments at t in such a location at such a time is itself necessary, it happens in every possible world, no matter what else happens. So, it could not possibly be responsive to the occurrence of any other event. There is no other world in which some other event e happens and you respond to e in a way that has you choose not to write the comment, for instance.

    I leave out consideration of counterpossibles. If we include them, then I think my case is closed. If you insist that you could have responded to e, since it is true that, had (counterpossily) e occurred, you would have so responded. Then I say that, by the same reasoning, it was metaphysically possible for you to do otherwise, since it is true that, had (counterpossibly) e occurred, it would have been metaphysically possible that you so responded.

    February 17, 2015 — 13:00
  • Leave a Reply to Michael Almeida Cancel reply

    Your email address will not be published. Required fields are marked *