In this short post I want to explain a more fundamental problem with Yujin Nagasawa’s problem of evil for atheists.
In my last post I granted, for the sake of argument, Nagasawa’s characterization of the problem of evil for atheists. Briefly, his idea is that it’s not only theists that have a problem of evil but atheists too. Theists have a problem of evil because, given God’s existence, they expect a very good world but it does not match the evil we see in the world. Most atheists have a problem of evil since most atheists are optimists–who think the world is not bad–and optimism does not match the evil we see in the world. This is the problem of axiological expectation mismatch — what we expect in terms of value (good or bad) does not match what we observe.
I think he misframes the problem for the atheist. For the theist the expectation is generated by the good God theory, and the observation is the unexpected evil we see in this world. But what does the atheist have as a predictive theory? Nagasawa seems to be treating optimism as theory, but that is odd because he points out:
[T]here is no straightforward link between atheism/nontheism and modest optimism because atheists/non-theists do not believe that we and our environment were created by an omnipotent and wholly good God (p. 145).
This is like saying people who believe they’re left-handed have a problem of evil, while acknowledging there is no conceptual link between lefties and the problem of evil. Most lefties may be optimists but the problem has nothing to do with that belief. If optimism is a predictive theory then it has as much to do with atheism as it has to do with lefties.
If anything, the predictive theory on atheism would be indifference, and optimism would be the observation. If this is right, then the atheist’s expectation is a morally neutral world (from indifference) and the observation is a not bad world. These atheist optimists actually have a problem of good, rather than evil.
I’ve been watching some youtube videos on fine-tuning lately, and I was thinking about this fairly old video where John Hawthorne was talking about Bayes’ and fine-tuning and how cheese would be evidence for a God with a cheese fetish. He says:
“I think the existence of cheese is evidence of an intelligent designer with a cheese fetish. I mean it is evidence, and I can use the same argument: the absence of cheese would surely be evidence against the existence of an intelligent designer with a cheese fetish, and it’s probabilistically incoherent to allow that the absence of cheese is evidence against without also allowing that the presence of cheese is evidence for. So it is evidence, but it’s such a ridiculous hypothesis to begin with, in that case, it might take you from one in a trillion to four in a trillion. That doesn’t do very much. So evidence just bumps you up, but it might not bump you to anything respectable in terms of taking something very seriously.”
For those familiar with how Bayes’ works, everything Hawthorne said seems reasonable. Since the cheese fetish God is such a ridiculous hypothesis, that would mean the prior is so low that, as he says, “it might take you from one in a trillion to four in a trillion.” This seems like a good result, but here’s where my worry comes in.
Imagine a possible world where we mirror our laws of physics. Suppose we know about this planet called Earth, and while there is no life on it, there is cheese. I think we can run a fine-tuning argument for a God with a cheese fetish, contrary to what Hawthorne seems to be implying. Hawthorne implies that since the prior on the cheese God is so low that the evidence of cheese won’t bump you up to anything respectable. In the following I’ll use the odds form of Bayes for a good God FTA and compare it to the cheese God FTA, and show that if the former argument is good then probably so is the latter despite the cheese God being “ridiculous”.
We’ll use the odds form of Bayes’ to run the fine-tuning argument.
So for the good God FTA, H1 will be the good God, H2 will be single universe naturalism, and E will be the evidence of fine-tuning. The ratio of posterior probabilities will tell us how much more likely the good God is over naturalism, given the evidence. For concreteness, let’s insert some toy numbers. Let P(E | good God) = 1 and let the ratio of the priors be 1. These toy numbers won’t make much difference because they will get swamped by the fine-tuning of the cosmological constant, which says that P(E | N) = 1/10^120. So, on these toy numbers the good God hypothesis is 10^120 times more likely than naturalism, given the evidence. The toy numbers would have to be extraordinarily off for naturalism to stand a chance.
Now, let’s run the FTA for a God with a cheese fetish. Let’s suppose the fetish is strong such that P(E| cheese God) = 1, and suppose the cheese God is “ridiculous” in the sense that the prior on the cheese God is a trillionth of the good God. (That’s a big difference!) One trillion is 10^12. When we plug in the “ridiculous” hypothesis the cheese God wins by 10^108. Again, these toy numbers would have to be extraordinarily off for naturalism to stand a chance.
Obviously, cheese can be a placeholder for anything mundane — like lint, soot or mud. Cheese (and soot and mud) also requires fine-tuning in that if the cosmological constant were not fine-tuned then the universe would not be cheese-permitting. In order for there to be cheese, we need stars to form and supernovas to happen. Yet it seems off to think that we should infer a designer with a cheese fetish based upon these cheese-permitting conditions.
I haven’t shown where the FTA goes wrong, but I think I have provided some reason to be suspicious of it, if it turns out that, given these laws of physics, we should infer design from such mundane things like cheese, soot, or mud.
While the problem of evil has typically been seen as a problem for traditional theism, Yujin Nagasawa has argued that it is also a problem for most atheists–namely those that think the world is not bad–and that theists have an advantage when it comes to solving this problem. So, since the problem of evil is more severe for atheism it provides “indirect motivation” to be a theist. This is a surprising result.
My initial reaction is that this argument can’t be right, but figuring out where it goes wrong is requires some work. I’ll briefly sketch the argument and give some pushback.
Nagasawa defines modest optimism as the view that “overall and fundamentally, the environment [Earth] in which we exist is not bad.” This seems to be a natural entailment of theism given that a good God created it. Indeed, Leibniz thought God had to create the best possible world. While modest optimism isn’t an entailment of atheism, most atheists are modest optimists. It is this modest optimism combined with the evil in the world that creates Nagasawa’s problem of evil. Nagasawa thinks the problem of evil is a specific version of a more general problem which he calls “the problem of axiological expectation mismatch.” The idea here is that what we expect in terms of value (good or bad) does not match what we observe. For example, in the problem of evil for theists, we expect there to be a very good world but it does not match the evil we see in the world. Similarly, atheist modest optimists expect there to be a world that is not bad but it does not match the evil we see. Nagasawa finds the evil of evolution and natural selection particularly troubling because the evil is ingrained in the biological system, where creatures must inflict pain on other creatures for their very survival. Unless you are a pessimist who thinks the world is overall bad, the problem of systemic evil is equally a problem for everyone.
Once it is clear that this is a problem for most atheists, Nagasawa’s next step is to argue that theists enjoy an advantage over atheists when it comes to solving this problem. The reason why theists have an advantage is that atheist ontology is a subset of theist ontology: in addition to the natural world, theists have the supernatural world. By analogy, suppose a builder is trying to build a house and he can choose among two toolboxes. Toolbox 1 has all the tools toolbox 2 has and some extra. Presumably if you want to finish the job, toolbox 1 has an advantage over toolbox 2. Theists have tools in their toolbox that atheists don’t have like soul-building and universal salvation, so theists have an advantage.This is the thrust of the argument.
It’s easy to overestimate the target of Nagasawa’s argument–it’s actually quite modest. He’s not trying to argue that theism is true–though he says it provides motivation for theism; He’s not trying to solve the problem of evil; He’s not claiming the world is better off with God. He’s merely arguing that–when it comes to maintaining modest optimism–theists have an advantage over atheists when it comes to solving the problem of systemic evil–a problem both sides have (mostly).
At first glance, this argument seems plausible, but upon closer inspection I think there is something important missing. The reason this argument seems plausible is that Nagasawa has–in an attempt to simplify the argument and subsume the widest audience into the definition of modest optimism–left out the fact that there are varying levels of expectation among modest optimists. Theists should expect a very good world, and perhaps the best world (as Leibniz thought). Atheist modest optimists, on the other hand, are unlikely to expect nearly as good a world. To put this into perspective, consider a scale between 0 and 100, with 50 being that the world is axiologically neutral. Modest optimists are anyone who assign a value of the world from 50 to 100. Suppose systemic evil leads atheists and theists to think the world is at 40; suppose atheist optimists expect the world to be 60; and suppose theist optimists expect the world to be 90. The expectation mismatch is only 20 for atheists, while it is 50 for theists. (I hope it is obvious that these are toy numbers, but it should nevertheless be instructive and a rough guide.) Nagasawa claims that if atheists find a naturalistic solution then theists can simply borrow the solution so it results in a draw–both sides have a solution. So let’s say there is some naturalistic solution that bumps the 40 up to 60 and theists borrow that solution. (Perhaps the solution is that the existence of humans is a greater good that is worth the cost of systemic evil.) Modest optimism is saved for both atheists and theists. Recall that Nagasawa says the systematic problem of evil is a specific version of the more general problem of axiological expectation mismatch. The atheist optimist now has no mismatch, while the theist still has a mismatch of 30. Strictly speaking, I think Nagasawa is correct: if the goalpost is merely to get over the 50 hump to save modest optimism then theists have an advantage–I do think this is Nagasawa’s goalpost. But if the goalpost is to collapse the mismatch, it is not enough for theists to borrow naturalist solutions for it to be considered a draw, since they likely start off with a bigger mismatch. Which is the more interesting goalpost? I say the latter, especially in the context of deciding between atheist modest optimism and theist modest optimism–a context that is not in the scope of Nagasawa’s argument. For if atheist modest optimist predictions are more accurate (less mismatch) than the theist modest optimist predictions, that would be evidence for atheist modest optimism.
Nagasawa generalizes the traditional problem of evil to includes atheists, pantheists, and axiarchists. In the spirit of generalization, we can borrow his strategy to say that for any expectation mismatch the theory with more ontology has an advantage. So–to pick a silly example–suppose I can’t find my car keys. There is an expectation mismatch from where I expect my keys to be and my observation that it’s not where I left it. Suppose theory 1 has everything theory 2 has but with the addition of Martians. Theory 1 has an advantage over theory 2. This is a fairly modest claim, because the defender of theory 1 is not saying that Martians probably took the keys. Rather, the mere possibility that Martians took the keys gives theory 1 an advantage over theory 2.
Ok, it’s more like a causal principle war, but I couldn’t pass up the title.
This post will be a follow-up of to my last post, where I will reply to Loke’s latest reply on the youtube channel Intellectual Conservatism. I will reply to some, but not all, of the issues raised in the video and also Loke’s written comments. I would recommend watching the youtube video before reading this, and maybe even switching between watching the video and reading my replies at the given timestamp.
Is Loke doing an internal critique?
At 6:16 Loke says that if Oppy’s theory entails a contradiction then it cannot be true, and this is what he shows with his Modus Tollens argument. In my previous comments section, Malpass asked Loke if he’s doing an internal critique, since, presumably, nothing interesting follows from the fact that Oppy’s theory contradicts Loke’s theory. Loke says yes, but it seems to me this is isn’t true. First, Oppy thinks the tu quoque to the MT argument is good, so it seems he thinks some premise of it must be false. Is Loke saying that Oppy accepts every premise in a valid argument but rejects the conclusion? This is not an interesting point, but Loke keeps bringing this up.
Loke’s statement of the problem
Here is how Loke states the problem from the comments in my previous blog:
We know of things/events/states which begin to exist e.g. my house, increasing in strength of electric field, stars, Big Bang. We know that whatever begins to exist could begin to exist (actuality imply metaphysical possibility=broadly logical possibility). So on Oppy’s view what is the relevant difference that explains why the initial state of reality (ISOR) begins uncaused but not others? As I explained in the debate, S is required. But S can only do the job of explaining why ISOR begins uncaused when ISOR has already begun uncaused. In that case, S is superfluous in accounting for why the actual initial state BEGUN rather than other broadly logically possible alternatives, because S comes too late, so to speak, in the order of explanation—the initial state already BEGUN to exists!
So Loke asks for a special property that metaphysically grounds why ISOR began to exist uncaused rather than other metaphysically possible alternatives. On an Aristotelian or branching theory of modality, all possible worlds share a necessary initial segment (in the context of the debate Oppy is granting a finite past). This means that any conceivable initial state other than ISOR is impossible, since there is only one possibility, namely, ISOR. Why ISOR rather than B? Because ISOR is necessary and B is impossible.
Here’s how this Aristotelian theory of modality works. Suppose the world is indeterministic and that a radioactive atom can decay at t1 or t2. This is to say that there are (at least) two possible ways the world can branch off to. These possibilities are grounded in the indeterministic causal powers of the radioactive atom. So the possibility of any later state of the world will depend on the prior state and its causal powers, just like how the atom’s decaying at t1 depended on the prior existence of the atom and its causal powers. We can think about how this goes until we go back in time to the first moment of time. Since the first moment of time has nothing causally before it (or so I say), there aren’t different ways initial state of reality could have gone; it’s necessary. There is no further fact for why ISOR is necessary, and it seems to me the same would have to be said for Loke’s view on why the Trinity is necessary rather than the Quadrinity. Here Loke would most likely reference his MT argument and say other things would begin to exist uncaused, but it seems to me that we should reject the metaphysical principles/premises that lead to this result, especially given that the tu quoque still seems fatal to me.
In explaining why ISOR began to exist uncaused rather than B, am I using the special property of necessity, or am I using a metaphysical principle? If I am using a special property, it doesn’t seem to me to be too late. I’m more inclined to say I’m using a metaphysical principle rather than a special property. Why think this metaphysical principle (or view on modality) is true? Here’s how Oppy argues for the Aristotelian view of modality in The Puzzle of Existence:
My favourite theory of modality has the evident advantage of theoretical frugality. On the one hand, if there are objective chances, then any theory of modality is surely committed to the possibility of the outcomes that lie in the relevant objective chance distributions. On the other hand, it is not clear that we have good reason to commit ourselves to any possibilities beyond those that are required by whatever objective chances there might be; at the very least, any expansion of the range of possibilities clearly requires some kind of justification.
So, as I understand it, he’s saying that contingent possibilities are explained by indeterministic causal powers. If you want to expand the range of contingent possibilities beyond those from indeterministic causal powers, this would require some justification.
The Too Close objection
At 9:45 Suan talks about the objection that if he wants to persuade his interlocutor he needs to appeal to common grounds and not use premises too close to the conclusion being rejected. I raised an objection like this in my comments section. Here’s how it goes. Recall that Loke uses his Modus Tollens argument to show that if we reject his causal premise—whatever begins to exist has a cause—then we get the bad result that other things would also begin to exist uncaused. Then, according to the tu quoque, Loke has the same problem because other timeless things would also exist uncaused. To defend against the tu quoque, Loke gives the symmetry-breaker of God being beginningless or initially timeless. One problem is that he uses a premise that’s very similar to the causal premise that’s being rejected; he’s saying something along the lines of “whatever begins to exist needs a special property … but timeless things don’t.” Why would anyone that rejects the causal premise accept this symmetry-breaker for the tu quoque? This symmetry-breaker is dialectically impotent given how close it is to the causal premise being rejected. Call this the Too Close objection.
Loke replies to the Too Close objection in two ways: (1) he thinks he does appeal to a common ground, and (2) what really matters is not common ground but whether his grounds correspond to reality. Concerning (1), to be clear, my Too Close objection is not that there’s no common ground; rather, it was directed at the use of timelessness in order to say that God doesn’t need a special property.
Abstract Objects
In my first post, I said we should use metaphysical principles rather than special properties, but Loke objected that abstract objects don’t make anything the case. He raised this objection numerous times in comments, so it should be addressed.
At 43:19 Loke responds to a tu quoque from Malpass: Doesn’t Loke also use abstract metaphysical principles like the causal principle? Loke replies:
I do refer to abstract principle ‘everything that begins to exist has a cause’, but this principle is just the consequence of my view that what makes things happen are concrete entities and their properties, which implies that without concrete causes doing the work nothing would begin to exist. That’s why uncaused events do not happen.
In Loke’s replies to me he also thought I was appealing to Platonic objects, but this is not the case. My metaphysical principles can have the same metaphysical status as his. Compare these two principles:
Loke: Whatever begins to exist has a cause. Oppy: Whatever is not the first cause has a cause.
I see no reason why the latter has to be Platonic, if the former needn’t be. Loke says the causal principle “is just the consequence of [his] view that what makes things happen are concrete entities and their properties.” It seems you could equally justify that Oppy’s principle is just the consequence of his view that what makes things happen are concrete entities and their properties, which implies that without concrete causes doing the work there wouldn’t be non-initial events. That’s why uncaused non-initial events don’t happen. After all, on the Aristotelian view of modality, modal truths are grounded in causal powers. Call this the Me Too objection (or me quoque?).
Two forms of the Modus Tollens
At 47:20 Loke talks about two forms of his MT argument. According to the first form, if Oppy’s ISOR began to exist uncaused then later events would also begin to exist uncaused. According to the second form, if Oppy’s ISOR began to exist uncaused then other initial events would have begun to exist uncaused. Either one is sufficient to debunk Oppy’s theory.
At 54:00 Loke says that Oppy’s view on modality is only relevant to the second form of the MT—the one about the initial state. This is because Oppy claims that the initial state is necessary, and there cannot be other possible spacetime block beginning to exist alongside ours. On the other hand, the first form of the MT concerns later events, which are actual events like the increase of the strength of an electric field. Loke’s reasoning seems to be that we know that it’s possible that electric fields increase in strength, because they actually happen; being actual entails that it’s possible, but not vice versa. And since there no relevant difference between these possible events and the Oppy’s ISOR, where beginning to exist uncaused is concerned, these later events would begin to exist uncaused.
It’s possible I’m not understanding the problem. Oppy’s view on modality seems relevant to both forms of the MT. The reason it’s relevant to the first form—concerning later events—is that what’s possible for these later events will be set by the indeterministic causal powers of the previous event. This is why an increase in the strength of an electric field doesn’t begin to exist uncaused. All later events are indeterministically caused on this theory of modality. There isn’t any later branch that isn’t set by a cause. (Note, the point here only concerns the relevance of Oppy’s theory of modality to the first form.)
At 56:42 Loke talks about the second form of the MT—the one concerning the initial state. Loke says Oppy’s theory of modality isrelevant to the second form, but it fails for two reasons. First, he says, Oppy’s theory of modality may be ok if the initial state is an immaterial, beginningless first cause, but it doesn’t work if initial state is a first cause which is a physical entity with beginning. He thinks this because physical things with beginnings can be arranged differently; for example, the tables and chairs in his room can be arranged differently, and, likewise, atoms can be arranged differently. So the thought is that it is problematic to think that the physical things of Oppy’s ISOR couldn’t be arranged differently, and hence be contingent. An immaterial, timeless first cause, on the other hand, doesn’t have this rearrangement problem.
It seems to me Oppy can simply deny this based on his preferred theory of modality. On that view, the possibility of some arrangement now doesn’t entail that it is possible as the ISOR. For example, while it’s possible that a rabbit begins to exist now, but that doesn’t mean that a rabbit is possible as the initial segment. What’s possible at any non-initial moment of time will depend on the previous state and its causal powers. It would take a long while in the evolution of the universe for rabbits to be possible. Contrary to what Loke says, we don’t know that the ISOR could be arranged differently through “science and observation.” We can observe physical things being rearranged at a specific time by its being caused, but the first cause is not something that can be caused to be rearranged in virtue of being first. The possible arrangements of later objects will depend on the previous state of the universe and its indeterministic causal powers. What Loke seems to be using here is a principle about modality. Maybe the principle is: Whatever can be rearranged is contingent. Oppy is free to reject this principle. Recall that Oppy favors his view based on theoretical frugality stating:
[I]t is not clear that we have good reason to commit ourselves to any possibilities beyond those that are required by whatever objective chances there might be; at the very least, any expansion of the range of possibilities clearly requires some kind of justification.
Loke is positing a contingent possibility that goes beyond those that come from causal powers.
Loke’s tu quoque symmetry-breaker of rearrangement
Recall that Loke thinks that Oppy’s branching theory of modality is ok for immaterial things but not for physical things, and this is because he has a symmetry-breaker, namely, we have no evidence that timeless immaterial things can be arranged differently, whereas we do for physical things. (I don’t think ‘arrangement’ or ‘rearrangement’ is the best word since that implies space, and presumably immaterial things, in this context, are non-spatial.) Loke thinks this symmetry-breaker dissolves the tu quoque. According to the tu quoque, if we follow Loke’s reasoning, then other timeless entities—like the Quadrinity—would exist. But, Loke says (I think), since we have no evidence that timeless immaterial things can be arranged differently, we have no evidence that they are contingent. Furthermore, Loke denies that conceivability entails metaphysical possibility. If he did accept it, I think his MT argument would easily fall to the tu quoque, since a Quadrinity is equally as conceivable as a Trinity. Now it’s going to take more work to run the tu quoque.
Loke says that we don’t know whether a Quadrinity is metaphysically possible. To support this, he says:
to prevent a quadrinity from existing uncaused beginninglessly in an initially changeless and timeless state, the preventing conditions can be part of the initially changeless and timeless state which makes such a state incompatible with a Quadrinity existing.
So the idea seems to be that, for all we know, the First Cause, in its timeless state, could have some preventing condition that prevents other timeless things, like the Quadrinity, from existing. To be clear, not only must it prevent the Quadrinity, but any other timeless thing. (I suspect he thinks this preventing condition is causal.) Given this epistemic possibility, Loke says the burden of proof is now on those that raise the tu quoque to “rule out” this epistemic possibility.
I don’t think that I have to rule it out. Isn’t it enough that a premise be more plausible than its denial, as Craig often says? So the question is: Is it more plausible than its denial that this timeless entity has this preventing condition? My initial thought is to wonder why Oppy’s ISOR can’t also have the preventing conditions be part of it such that it makes the alternatives incompatible with it; the epistemic possibility applies to Oppy’s ISOR just as well (another me quoque). In reply, Loke says we need to distinguish between states and events. For events the preventing conditions must be prior to it, but for timeless states, the preventing conditions can be part of it. I find it mysterious why this is so. Why is the distinction between events and states relevant? Causal priority can apply to the timeless state as well. Ask modified theistic activists who think God causes the timeless Platonic horde. And what special causal powers or properties can timeless things have that temporal things can’t? Loke merely offers the epistemic possibility of some preventing condition without specifying what that could be. Maybe Loke doesn’t know either what it could be. Maybe it’s simply a primitive power to prevent other things from existing. I see no reason why timeless entities have the advantage here.
Loke’s tu quoque symmetry-breaker of beginninglessness
At 1:16:28 they take about the beginninglessness symmetry-breaker. Recall the original tu quoque that Loke’s reasoning behind his MT would lead to other timeless entities existing. These special properties come “too late” to metaphysically ground why one thing begins to exist rather than another, and, according to the tu quoque, this applies to God’s timeless existence as well. Loke offers the property of beginninglessness as a symmetry-breaker for why God doesn’t need a special property. He replies:
It should be noted that I am NOT claiming that anything that exists requires a special property to explain why it exists. Such a principle is obviously false. E.g. my existence does not require a special property to explain why I exist. Rather, my existence is explained by my already-existing parents who brought me into existence and I am not required to have a special property S. …
God is supposed to have always already existed at all earlier durations and in the initial timeless state (which is what is meant by beginningless), hence no already-existing preexisting causes are required and He is not required to have a special property S. In this case, His beginninglessness is not a special property that explains why He exists; rather His beginninglessness is merely a way of describing His always-already existence, which also imply that no causes are required i.e. He would be uncaused. Hence, God exist uncaused doesn’t need to be explained by S.
It might asked whether ‘beginningless’ itself is a special property S. in reply, there are two distinct senses of explanation which needs to be clarified (1) a statement or account that makes something clear (OED). (2) to provide a metaphysical grounding for. In the case of ISOR begins uncaused, I was arguing that there needs to be a special property S that not only makes something clear but also provides a metaphysical grounding for why ISOR begins uncaused but B begins caused (but there cannot be such an S). In the case of God existing beginninglessly, God’s beginninglessness merely makes clear why is it the case that no cause or special property is needed (it doesn’t provide any metaphysical grounding which S is supposed to provide). In particular, by explicating the meaning of beginninglessness, we can see why it implies that God would be uncaused (assuming that God and ISOR are both unsustained). Thus, ‘beginningless’ itself is not a special property S.
In his book, Loke writes:
Proponents of the Kalam would argue that God being uncaused can be explained by God being beginningless (see Chap. 6), in which case being beginningless is a non-causal explanation and a property that is possessed by God when God already exists. (2017, p. 148)
In the book he says beginninglessness is a non-causal explanation for God being uncaused even though it is a property that God has when he already exists. In his above reply he clarified that he’s using the OED definition of “a statement or account that makes something clear,” and not a metaphysical grounding explanation. God’s being beginningless OED-explains why no special property is needed, and this implies that God is uncaused. To make it consistent with his book, the ‘implies’ must also be an OED-explanation, so God’s being beginningless also OED-explains why God is uncaused. We have two OED-explanations here: one for why no special property is needed and one for why God is uncaused. I take an OED-explanation to describe any kind of explanation since the point of any explanation is to make things clear. There are many kinds of explanations—grounding, causal, personal, mathematical etc.—and they all fit the OED explanation. I wonder what kind of explanation Loke is using to explain God’s existing uncaused, because it’s only fair that Oppy can use the same kind. Loke has already stated that it’s not a grounding explanation, and there can’t be a causal explanation to the first cause, so maybe he’s using metaphysical principles.
This part of the dialectic is a little bit of a tangent though, because OED-explaining why God exists uncaused by being beginningless is not relevant the tu quoque. We want to know why God exists uncaused rather than all the other metaphysically possible timeless beings. This is parallel to the challenge Loke gave to Oppy. Being timeless could also explain why a Quadrinity exists uncaused, and if there is no special property that the Trinity has, then—by the tu quoque—the Quadrinity would also exist uncaused. Since I already went over Loke’s objection to the Quadrinity objection, I’ll move on.
Edit 12/28: It occurred to me I missed a point about the Quadrinity. Loke says the burden of proof is on those raising the tu quoque to prove that the Quadrinity is metaphysically possible. In one sense this is easy, and in another sense this is hard. In his book (chapter 5) he talks about both the logical and metaphysical possibility of an entity/state of affairs. In his first debate with Oppy on Capturing Christianity, he also referred to the various logical possibilities of the universe. If he is going to use logical possibilities to run the MT on Oppy, then it’s fair to use logical possibilities to run the tu quoque back on him. By logical possibilities, he says he means the things that do no violate the laws of logic. In this sense, it’s easy to run the tu quoque since the Quadrinity or timeless ghosts don’t violate the laws of logic. It is after this first video where I ran the tu quoque. Things get tougher with a shift to metaphysical possibility, for one’s modal epistemology is important here. (It seems we’ve shifted to metaphysical possibility, and talk of logical and factual necessity has disappeared.) On one view, some suitably quaified version of conceivability entails possibility. This is one of the easier ways that would show that the Quadrinity is metaphysically possible, since the Quadrinity seems equally conceivable as the Trinity. But Loke says, “what’s conceivable is not always possible,” so it seems he rejects this principle. If we’re running Loke’s more austere modal epistemology (though he hasn’t specified it in detail), I suspect it’s going to be hard to prove the Quadrinity’s metaphysical possibility. Summing up: the tu quoque is easy if we’re using logical possibility, or conceivability, and hard if we’re using some austere form of modal epistemology. At the same time, the more austere one’s modal epistemology is, the harder it is to show that there are more possibilities than Oppy’s ISOR, which undermine the original MT.
This ends my reply to the video, but I want to conclude by raising two more points: another me quoque objection, and possible ramifications for theistic arguments.
Another me quoque objection
In the original debate, Loke says that the First Cause must be initially changeless. I can’t make sense of Craig’s hybrid view of God and time. On this view, God is timeless sans creation and temporal since creation. Supposing the view makes sense, why can’t the naturalist use it too? In the debate, Oppy pointed out that “nothing changes at a point.” So consider the first moment of time at t=0. Isn’t that initially changeless? (Isn’t every instant of time changeless?) Change happens when a substance gains or loses a property, and this takes duration. Is there any change at t=0? No. So, contrary to what was originally stated, the initial state of the universe isn’t at t=0 but timeless, since it’s initially changeless. We can now say that the universe is timeless sans change, and temporal since change such that it is not susceptible to Loke’s causal principle.
Possible ramifications for theistic arguments
I think there are some theists that will want to reject the reasoning behind Loke’s Modus Tollens argument. Perhaps you’re a theist that thinks God has the special property of perfection that metaphysically grounds why he exists rather than some other beginningless being. Perhaps you’re modified theistic activists like Davis and Gould that think that God creates the Platonic horde of abstract objects. If God is uncaused, what special property does God have that Platonic objects don’t? Since there can be no special property due to the “too late” objection, these Platonic objects would be uncaused, contrary to modified theistic activists.
Recently Oppy and Loke had a debate on the Kalam. In this post I want to point out some flaws in Loke’s argument. One difference between Loke’s Kalam and Craig’s is that Loke has what he calls the Modus Tollens argument. Here’s how Loke states the argument at 50:17:
If our universe begins to exist uncaused, other things would also begin to exist uncaused. Why? Because, firstly, there will be no cause that makes it the case that only the universe rather than other things begin to exist uncaused. Secondly, whatever properties that differentiate the universe and other things will be had by these things only when they had already begun. Thirdly, the circumstances is compatible with other things beginning to exist. So, given 1, 2, and 3, this implies there will be no difference between our universe and other things and other events where beginning to exist uncaused is concerned. So if there is no relevant difference what this means is that other things would also begin to exist uncaused.
It is not the case that those other things begin to exist uncaused.
It is not the case that our universe began to exist uncaused.
Loke’s idea is that if Oppy rejects the causal premise of the Kalam—whatever begins to exist has a cause—then “those other things would also begin to exist uncaused.” Since Oppy doesn’t think the non-initial things begin to exist uncaused, he accepts Loke’s 2), and this means he denies Loke’s 1).
This implications of rejecting the causal premise goes at least back to Craig, as he writes:
If things really could come into being uncaused out of nothing, then it becomes inexplicable why just anything or everything does not come into existence uncaused from nothing. Why do bicycles and Beethoven and root beer not pop into being from nothing? Why is it only universes that can come into being from nothing? (Craig and Sinclair 2009, p. 186)
Craig formalizes the argument like so:
If it is possible for something to come into being without a cause at a first moment of time, then it is possible for things to come into being without a cause at later moments of time.
It is not possible for things to come into being without a cause at later moments of time.
Therefore, it is not possible for something to come into being without a cause at a first moment of time (Craig 2010).
But Loke wants to make a bolder claim than Craig. He changes Craig’s 1) to:
If the initial state of reality began to exist uncaused, then certain states of affairs would begin to exist uncaused at later moments of time.
So while Craig is saying if we reject the causal premise of the Kalam, then it is possible for things to pop into existence at later times, Loke is saying that it’s guaranteed that things to pop into existence at later times. Loke’s reasoning, as he states in 1), is that if there’s no relevant difference between Oppy’s actual initial state and other logically possible initiate states, then if the former pops into being then the latter will too. (Presumably the space of other logically possible states is actually infinite, so the universe just got real crowded!)
The way out for Oppy, Loke says, is to posit some special property S that the actual initial state has that other logically possible initial states don’t have that metaphysically grounds (Loke 2017, p. 146) why the actual initiate state occured rather than the alternatives. (‘Metaphysical grounding’ is often used liberally and can mean a lot of different things, but in this context it has to be some kind of non-causal explanation. It can’t be a causal explanation since we’re talking about the First Cause, and it’s analytic that there’s no cause prior to the first cause.) But before you answer what the special property S is, there’s a problem, says Loke. In order for the initial state to have S, it would already have to exist. In that case, S is superfluous in accounting for why the actual initiate state occurred rather than other logically possible alternatives, because S comes too late, so to speak, in the order of explanation—the initial state already exists! To sum up, Loke originally asked for a special property S that would explain why the actual initiate state occurred rather than the alternatives, but then it turns out that it doesn’t matter a lick what S is, since S arrives too late for it to do any explaining. Loke’s conclusion: it’s impossible to explain why the actual initial state occurred rather than alternatives, if we reject his causal premise. Furthermore, given that there’s no special property explaining the relevant difference between the actual initial state and other logically possible states, the universe just got really crowded with things guaranteed to pop into existence. That’s how Loke sees it anyway. I don’t buy this reasoning, but let’s continue.
So the problem is that Oppy can’t explain why one initial state occurred uncaused rather than an alternative one. It’s natural to wonder how we could explain God’s existing uncaused without falling into similar problems. After all, whatever special property S that we use to explain why God is uncaused will already have to be had by God, so that S should be similarly superfluous. Loke is aware of this worry, so how does Loke reply?
The requirement for explaining ‘why something begins to exist uncaused’ is different from the requirement for explaining ‘why something is uncaused’. In the case of … [God] … ‘beginningless’ is a property that can do the explanatory job for why God exists uncaused … (2017, p. 148-9).
So Loke makes a distinction between a) beginning to exist uncaused and b) being uncaused. In the case of the latter, it’s ok to use the special property of being beginningless. I can’t make sense of this. How is the difference between a) and b) relevant to the problem? It’s still the case that God would already have to exist to have that special property. This same reasoning was used to block Oppy’s reply but Loke doesn’t apply it to himself. This seems like special pleading.
One might object that that this is merely a tu quoque on my part; I’ve only said the same problem applies to Loke too without showing where his argument goes wrong. That’s true, but my tu quoque is useful because if the same problem afflicts Loke’s view, then that gives us no reason to favor his view over Oppy’s. But I think I can do one better; I can show where the argument goes wrong. The problem is in having this requirement of a special property S that (in some impossible sense) exist before the thing in question. Instead of appealing to a special property that things already have (whether timelessly or temporally), we appeal to a timeless, metaphysical principle.
In fact, this seems to be just what Oppy does, and what Loke should do. Loke has his principles, including the causal premise, and Oppy has his principles: 1) everything has an explanation, 2) all non-initial things have causes. Both Loke’s and Oppy’s principles would explain why non-initial things don’t pop into existence uncaused. So Oppy says that if we compare theories, his theory is simpler and explains the same things in the universe. So he wins on simplicity: he has one less entity. This is Oppy’s strategy: comparing theories and seeing whose has the better theoretical virtues.
While this is fine, I think we can do better. I think we can do an internal critique of Loke’s view. I’ve already mentioned the special pleading, but there is one more case of potential special pleading. Recall the first premise of Loke’s Modus Tollens argument. Here’s my parody. I keep the structure of his premise while replacing a few words:
If God exists uncaused, other timeless things would also exist uncaused. Why? Because, firstly, there will be no cause that makes it the case that only the God rather than other timeless things exist uncaused. Secondly, whatever properties that differentiate God and other timeless things will be had by these things only when they already exist. Thirdly, the circumstances is compatible with other timeless things existing. So, given 1, 2, and 3, this implies there will be no difference between God and other timeless things. So if there is no relevant difference what this means is that other timeless things would also exist.
So if we accept the reasoning behind Loke’s first premise and apply it to God, we get the result that all the logically possible timeless beings would exist, a seemingly actual infinity. Timeless beings could be made of ectoplasm or perhaps they’re all immaterial minds. Surely an unwelcome result.
I conclude with two points as a reminder: 1) Loke special pleads by using a special property that God can only have if he already exists, while Oppy isn’t allowed to, and 2) if we keep Loke’s reasoning behind the first premise of his Modus Tollens, then we have a similar unwelcome result with timeless beings.
Edit 12/10: It occurred to me that I should add one more of Oppy’s principles to address one of Loke’s worries. Oppy holds to a branching view of possible worlds, where all possible worlds share an initial state, and he holds this view because it is parsimonious (see his chapter in Puzzle of Existence). So when Loke asks, “Why is the initial state A rather than B?” Oppy could answer, “Because A is necessary, and it follows from my branching view on modality.”
Hugh Jidiette 