One thing I enjoy about your posts is that you write remarkably clearly and consider plenty of reasonable objections. It's very nice to read these kinds of posts.
I much enjoyed this passage:"Usually when making arguments for (or against) theism based on what God would do (such as fine-tuning), the moral facts are held more or less in the background. So we attempt to figure out what God would do based on what our moral intuitions say, and then see how well this matches what we find (in rough terms). But if what I’ve argued is correct, this is much too simplistic. After all, 100% of our credence given theism shouldn’t be given to God acting in accordance with our moral intuitions. Rather it should also be distributed across alternative contents of morality, which would predict our intuitions being this way. "
I share your more than reasonable doubts in God saving moral knowledge.
On the other hand the possibility of the existence of a God is a great idea to get us humans to act well (memetically adaptive) we should be careful in throwing the meme away. But this is talking about something else entirely.
Thank you very much for the kind words! I was actually worried that I didn't communicate my thoughts very clearly in this post, as my grasp of the issue is still quite muddled, so I really appreciate it!
But are my intuitions about what God would want *strictly* moral ones? Bentham's Bulldog has made the point (quite recently: https://benthams.substack.com/i/158737265/might-there-be-an-evil-designer) that God's goodness is just emergent from His ultimate knowledge and power in the world. It seems unlikely that a consciousness creating consciousnesses would want those consciousnesses to be constantly terribly confused about things *in general.* So we should expect God to give us knowledge (or the capacity to develop knowledge) about all sorts of things—physics, math, modalities, and morality—which is what we find. I don't think God wants me to know what is good because it's good for me to know it—I think He wants me to know because it'd be super weird if he didn't, in a sort of epistemic sense.
I'm worried that this might create another analogous circularity—if I think God wants me to know true things because He likes it when people know true things because true things are true, then why is it I think that? Presumably, my intuition about His nature is being explained *by* His nature, so it all unravels again, but in a full-on external world skepticism kind of way. Of course, it seems like we have slightly stronger not-God reasons to deny external world skepticism (which you've written about), so maybe those arguments can push the God-truth circularity to a higher probability from the outside? Which then lets you make moral knowledge-y arguments for God as a subclass of the knowledge-at-all-y arguments.
I'm very unsure about all this, extremely interesting post!
That's an interesting approach! I think you're right that it just devolves into an analogous problem, but concerning God's reasons overall, rather than just his moral reasons, and it seems like all the same points apply.
As to whether the regular external-world skepticism responses save this regress I'm not so sure. Whatever things we can point to, without reference to God, that defeat external world skepticism, will be available to the atheist as well (I presume). So all the knowledge I can have without God will be exempt from this argument.
If we then zoom in on the subset of knowledge that requires God (such as moral knowledge, and presumably knowledge of his other reasons), the justification for that will run orthogonally to whatever justification you have for the non-God knowledge. After all, that knowledge will be justified through mediation of God (and thus through his actions), whereas the non-God knowledge will be "immediate." But then so long as the non-God knowledge doesn't include knowledge about what reasons God would have if he were to exist (which I don't think the response to external world skepticism provides, at least), the problem for moral knowledge persists.
The question is then simply whether the non-moral reasons you point to can be known on atheism, and if so whether that's sufficient to save the moral knowledge. I'm not sure about either, but prima facie there seems to be no relevant difference between the moral and non-moral reasons.
I'm not sure whether my reasoning here makes sense, or whether it's any good--I too am very unsure.
Great read! Here’s where I disagree (sorry for how sloppy this will be, walking home rn): the moral argument (or at least, by my lights, the most plausible version of it) doesn’t start by *assuming* that our moral intuitions aren’t evidence of the content of morality. Instead, the thought is that while our moral intuitions do give us evidence of certain moral propositions (simply in virtue of their appearing plausible to us), nevertheless naturalism provides a *defeater* of this reason, because it either defeats or positively undercuts the reliability of our moral intuition. Comparatively, theism does not give us any *positive reason* to doubt the prima facie reliability of our moral intuition.
Now, what you say is true — it may well be that the moral facts are different than how they appear to us, and God might have reason to deceive us (and, you are right that we cannot invoke our intuitions as evidence against that possibility). But that isn’t a positive reason to doubt the reliability of our moral intuition. In the same way that the possibility of an evil demon isn’t positive reason to doubt the reliability of our perceptual faculties. It is definitely possible we are wrong in various domains, but that isn’t the same as providing some positive debunking story on which we should *positively expect*, on a given theory (naturalism), that our intuitions are unreliable. Basically: the moral argument purports to show that, on naturalism, we should expect our moral intuitions to be unreliable (direct defeater). On theism, there is a mere absence of any such defeater (which is consistent with everything you said in this piece).
> naturalism provides a *defeater* of this reason, because it either defeats or positively undercuts the reliability of our moral intuition. Comparatively, theism does not give us any *positive reason* to doubt the prima facie reliability of our moral intuition
I don't think there is any difference here.
On naturalism our moral intuitions are the result of evolution through natural selection. They could just so happen to correspond to the True Morality (supposing that it actually exists), but there is no reason to expect that.
On theism our moral intuitions are the result of what God's want our moral intuitions to be. They could just so happen to correspond to True Morality (supposing that it actually exists), but there is no reason to expect that.
What you’re missing in your outline is the epistemic structure.
We start with our intuition (our seemings) — that it is morally wrong to boil a child. This gives us reason to think (and if it is true, knowledge) that it is wrong to boil a child.
Now, suppose I’m a naturalist. Well, then I realize that this faculty was created purely by evolutionary process: an evolutionary process that is insensitive to the actual moral facts. If there were no moral facts, nevertheless the same physical events would’ve transpired and I would end up with this faculty and intuition. Thus, there is literally no reason to think (and likely positive reason to think it’s not the case) that the faculty tracks morality, because it was totally uninfluenced by it in its creation. Notice that this is a true defeater: it does not make reference to a merely possible scenario. It makes reference to actual truths: the insensitivity of the evolutionary process to moral facts.
Now, suppose I’m a theist. What’s the defeater of my seeming that it is wrong to boil a child? You and this post say: the moral facts *might* be such that they direct God to deceive us. And, while we have a direct intuition that this is not the case, that cannot refute the possibility, since if God were deceiving us we’d have that intuition. This is all true. And yet, that doesn’t amount to a defeater. Because we have a seeming that the moral facts are such that it’s good for us to have moral knowledge. The mere possibility that the moral facts might be different isn’t a positive reason to think this seeming is false or unreliable. Just like the mere possibility of external world deception isn’t a positive reason to think our physical seemings are false or unreliable.
TLDR; the way you set it up is false: “there is no reason to expect God to set up our intuitions according to true morality”. There is: we’ve got a positive seeming that moral knowledge is good. That is not rebutted or undercut by the mere fact that it is possible that God is deceiving us.
Correct me if I'm wrong. What you are saying is that there is no plausible story on naturalism according to which our moral intuition correspond to Objective Moral Facts (OMF), other than pure coincidence
On the other hand, on theism there is a plausible story according to which our moral intuitions correspond to OMF: God set up our evolution specifically this way to make our intuitions to correspond to moral facts because having moral intuitions corresponding to OMF is good and it's therefore represented in our moral intuitions.
I think I agree with that. Now, if probability of this story was itself higher than pure coincidence, that would be a point in favor of theism in regards of probability of objective morality.
But this doesn't appear to be the case! There is no particular reason why this story is more likely to be true compared to all the other possible stories. In other words, even though on theism we can reduce the improbability of us being right about OMF regarding boiling babies to the improbability of us being right about OMF regarding having correct moral beliefs, the latter is just as improbable as the former so no reduction in improbability happens.
> Just like the mere possibility of external world deception isn’t a positive reason to think our physical seemings are false or unreliable.
It is, though. It's prudent to consider this possibility and I think that initially some sort of subjective idealism is the correct ontological stance, given our epistemological situation. But then, as it's mentioned in the post:
> The thing about the external world hypothesis is that it compresses a lot of highly specific predictions (that turn out true) into a much simpler theory than alternatives. With moral realism I just don’t see why the moral facts being the way we think they are predicts our moral intuitions much better than deviant facts, given theism.
That’s not my claim. I’m thinking about it in terms of *defeaters*. I am not making the claim that I have some intuition-independent reason to think the theistic story on which we have moral knowledge is more probable than the deceptive stories. What I’m claiming is that our moral intuitions are defeated by naturalism, but not so by theism.
We start with moral intuitions about moral facts and the question is whether they have been defeated or not by something. On naturalism, we have a defeater. On theism, we do not. You’re right that we have no “3rd person” reason to think any of the theistic stories is more likely than the other. But it is mistaken to take this to be a defeater of your seeming that some moral fact is true — indistinguishable possibilities aren’t defeaters (and I bolster this by pointing out the analog of this point in the external world skepticism case). You respond by saying that it *is* a defeater in the external world context, and it’s just that the skeptical hypothesis is itself intrinsically less likely (because it is less parsimonious). But this strikes me as a clearly mistaken response in the context of external world skepticism, for the usual reason mentioned in the literature — the demon can simply be feeding you the seeming that parsimony is relevant for probabilistic determinations. You might respond that we must start with *some* principles to evaluate hypotheses, but the moment we do so, you’ve granted that indistinguishability based on one’s current evidence doesn’t constitute a defeater. Additionally, it’s just not at all clear to me that all skeptical hypotheses really are intrinsically less likely (they can posit significantly less entities, etc.)
I noticed that you are talking in terms of "defeaters", but I don't see how this term is useful. When I'm trying to translate your reasoning to the terms of conditional probabilities it really seems that it's not. Whether we have a "defeater" or not, probability of objective morality is still the same - as low as a simple coincidence. I'm not saying that in your terms we have a "defeater" of objective morality on theism, I'm saying that regardless of it it our reasoning about probabilities is the same.
So am I wrong here? Or is this "defeater" some completely different thing that does not affect probabilities in this case and our disagreement is that you believe that one should be reasoning about epistemology not just in terms of probabilities but also using this extra entity as well?
> it’s just that the skeptical hypothesis is itself intrinsically less likely (because it is less parsimonious)
A priori we can start from considering both hypothesises equally likely. But then as we collect more evidence about our reality, naturalism gets boosted by the fact that our world seem to be working on strict laws. There is no particular reason why would a demon make my perception of the world ordered like this, but there are all the reasons for it on naturalism.
We can update our deciever demon hypothesis to accomodate the evidence, saying that the demon just so happens to want the world to appear ordered, but this makes the hypothesis more specific and therefore less probable - only some part of probability mass of all possible deciever demons lies withing this more specific hypothesis.
> But this strikes me as a clearly mistaken response in the context of external world skepticism, for the usual reason mentioned in the literature — the demon can simply be feeding you the seeming that parsimony is relevant for probabilistic determinations.
It can, but whether a particular epistemical stance systematically produces truth is an empirical question. Try thinking that all things being equal the more complicated hypothesises is more likely to be true and see how well it works for you.
> Additionally, it’s just not at all clear to me that all skeptical hypotheses really are intrinsically less likely (they can posit significantly less entities, etc.)
Care to bet in advance which parts of the world are compressed like that? So far it really seems that the universe is bruteforcing everything and not saving any computations at all.
The important point here is that we can actually make accurate predictions about the behaviour of reality. We consult our naturalistic model and it systematically works very well.
To match this, we have to postulate that the demon just outright created the perfectly coherent simulation, including which epistemical stances are rational, that works exactly as if it was a naturalistic universe. And at this point our sceptical hypothisis becomes indistinguishable from naturalistic hypothesis + a demon did it. Which is just fine? The demon part is weird and doesn't fit our usual epistemology, but hey we can register it as a bold prediction, that our epistemic techniques stop working at this particular instance, which itself is very unlikely to the best of our knowledge, but still possible. Essentially it all added back up to normality.
>”Try thinking that all things being equal the more complicated hypothesises is more likely to be true and see how well it works for you”
This doesn’t seem to engage with the point I made — that our taking this stance wrt hypotheses seems to consistently work out can be explained perfectly by the posited scenario, namely, that the demon feeds you this seeming. You need to simply start with some principles about how to evaluate hypotheses.
>”I'm not saying that in your terms we have a "defeater" of objective morality on theism, I'm saying that regardless of it it our reasoning about probabilities is the same”
The structure of defeaters is really crucial to the point I am making. The reason is because I think we all *need* to have the following picture: there are some beliefs we have that are justified (and thus if true basically constitute knowledge) in virtue of the fact that we have a seeming in their favor that is not subject to any defeaters. Now, it’s a tricky question whether this commitment to defeaters amounts to a difference in a Bayesian analysis — that’s something debated among epistemologists and I don’t exactly know where I land. But what I’ve always thought is that it doesn’t matter, at least for purposes like ours here.
The reason is this: if the structure of defeaters does make a difference to the conditional probabilities of hypotheses, it’s best understood as some sort of analysis of prior probabilities. If not, the relevant point is the conditional (if seems that X & no defeaters —> justified in believing X) still needs to be true for one to be justified in, e.g., believing in the external world in light of skeptical scenarios. As long as the conditional is *true*, I can make the point I made in my original comment. Other approaches to skepticism inevitably need to pull on a principle like this in order to justify the standards by which they’re evaluating hypotheses in the first place (this is the first point I made in this comment).
One thing I enjoy about your posts is that you write remarkably clearly and consider plenty of reasonable objections. It's very nice to read these kinds of posts.
I much enjoyed this passage:"Usually when making arguments for (or against) theism based on what God would do (such as fine-tuning), the moral facts are held more or less in the background. So we attempt to figure out what God would do based on what our moral intuitions say, and then see how well this matches what we find (in rough terms). But if what I’ve argued is correct, this is much too simplistic. After all, 100% of our credence given theism shouldn’t be given to God acting in accordance with our moral intuitions. Rather it should also be distributed across alternative contents of morality, which would predict our intuitions being this way. "
I share your more than reasonable doubts in God saving moral knowledge.
On the other hand the possibility of the existence of a God is a great idea to get us humans to act well (memetically adaptive) we should be careful in throwing the meme away. But this is talking about something else entirely.
Thank you very much for the kind words! I was actually worried that I didn't communicate my thoughts very clearly in this post, as my grasp of the issue is still quite muddled, so I really appreciate it!
Cool intellect, dude, seems very giant...
But are my intuitions about what God would want *strictly* moral ones? Bentham's Bulldog has made the point (quite recently: https://benthams.substack.com/i/158737265/might-there-be-an-evil-designer) that God's goodness is just emergent from His ultimate knowledge and power in the world. It seems unlikely that a consciousness creating consciousnesses would want those consciousnesses to be constantly terribly confused about things *in general.* So we should expect God to give us knowledge (or the capacity to develop knowledge) about all sorts of things—physics, math, modalities, and morality—which is what we find. I don't think God wants me to know what is good because it's good for me to know it—I think He wants me to know because it'd be super weird if he didn't, in a sort of epistemic sense.
I'm worried that this might create another analogous circularity—if I think God wants me to know true things because He likes it when people know true things because true things are true, then why is it I think that? Presumably, my intuition about His nature is being explained *by* His nature, so it all unravels again, but in a full-on external world skepticism kind of way. Of course, it seems like we have slightly stronger not-God reasons to deny external world skepticism (which you've written about), so maybe those arguments can push the God-truth circularity to a higher probability from the outside? Which then lets you make moral knowledge-y arguments for God as a subclass of the knowledge-at-all-y arguments.
I'm very unsure about all this, extremely interesting post!
Thanks, yours doesn't seem too small itself!
That's an interesting approach! I think you're right that it just devolves into an analogous problem, but concerning God's reasons overall, rather than just his moral reasons, and it seems like all the same points apply.
As to whether the regular external-world skepticism responses save this regress I'm not so sure. Whatever things we can point to, without reference to God, that defeat external world skepticism, will be available to the atheist as well (I presume). So all the knowledge I can have without God will be exempt from this argument.
If we then zoom in on the subset of knowledge that requires God (such as moral knowledge, and presumably knowledge of his other reasons), the justification for that will run orthogonally to whatever justification you have for the non-God knowledge. After all, that knowledge will be justified through mediation of God (and thus through his actions), whereas the non-God knowledge will be "immediate." But then so long as the non-God knowledge doesn't include knowledge about what reasons God would have if he were to exist (which I don't think the response to external world skepticism provides, at least), the problem for moral knowledge persists.
The question is then simply whether the non-moral reasons you point to can be known on atheism, and if so whether that's sufficient to save the moral knowledge. I'm not sure about either, but prima facie there seems to be no relevant difference between the moral and non-moral reasons.
I'm not sure whether my reasoning here makes sense, or whether it's any good--I too am very unsure.
Great read! Here’s where I disagree (sorry for how sloppy this will be, walking home rn): the moral argument (or at least, by my lights, the most plausible version of it) doesn’t start by *assuming* that our moral intuitions aren’t evidence of the content of morality. Instead, the thought is that while our moral intuitions do give us evidence of certain moral propositions (simply in virtue of their appearing plausible to us), nevertheless naturalism provides a *defeater* of this reason, because it either defeats or positively undercuts the reliability of our moral intuition. Comparatively, theism does not give us any *positive reason* to doubt the prima facie reliability of our moral intuition.
Now, what you say is true — it may well be that the moral facts are different than how they appear to us, and God might have reason to deceive us (and, you are right that we cannot invoke our intuitions as evidence against that possibility). But that isn’t a positive reason to doubt the reliability of our moral intuition. In the same way that the possibility of an evil demon isn’t positive reason to doubt the reliability of our perceptual faculties. It is definitely possible we are wrong in various domains, but that isn’t the same as providing some positive debunking story on which we should *positively expect*, on a given theory (naturalism), that our intuitions are unreliable. Basically: the moral argument purports to show that, on naturalism, we should expect our moral intuitions to be unreliable (direct defeater). On theism, there is a mere absence of any such defeater (which is consistent with everything you said in this piece).
> naturalism provides a *defeater* of this reason, because it either defeats or positively undercuts the reliability of our moral intuition. Comparatively, theism does not give us any *positive reason* to doubt the prima facie reliability of our moral intuition
I don't think there is any difference here.
On naturalism our moral intuitions are the result of evolution through natural selection. They could just so happen to correspond to the True Morality (supposing that it actually exists), but there is no reason to expect that.
On theism our moral intuitions are the result of what God's want our moral intuitions to be. They could just so happen to correspond to True Morality (supposing that it actually exists), but there is no reason to expect that.
What you’re missing in your outline is the epistemic structure.
We start with our intuition (our seemings) — that it is morally wrong to boil a child. This gives us reason to think (and if it is true, knowledge) that it is wrong to boil a child.
Now, suppose I’m a naturalist. Well, then I realize that this faculty was created purely by evolutionary process: an evolutionary process that is insensitive to the actual moral facts. If there were no moral facts, nevertheless the same physical events would’ve transpired and I would end up with this faculty and intuition. Thus, there is literally no reason to think (and likely positive reason to think it’s not the case) that the faculty tracks morality, because it was totally uninfluenced by it in its creation. Notice that this is a true defeater: it does not make reference to a merely possible scenario. It makes reference to actual truths: the insensitivity of the evolutionary process to moral facts.
Now, suppose I’m a theist. What’s the defeater of my seeming that it is wrong to boil a child? You and this post say: the moral facts *might* be such that they direct God to deceive us. And, while we have a direct intuition that this is not the case, that cannot refute the possibility, since if God were deceiving us we’d have that intuition. This is all true. And yet, that doesn’t amount to a defeater. Because we have a seeming that the moral facts are such that it’s good for us to have moral knowledge. The mere possibility that the moral facts might be different isn’t a positive reason to think this seeming is false or unreliable. Just like the mere possibility of external world deception isn’t a positive reason to think our physical seemings are false or unreliable.
TLDR; the way you set it up is false: “there is no reason to expect God to set up our intuitions according to true morality”. There is: we’ve got a positive seeming that moral knowledge is good. That is not rebutted or undercut by the mere fact that it is possible that God is deceiving us.
Correct me if I'm wrong. What you are saying is that there is no plausible story on naturalism according to which our moral intuition correspond to Objective Moral Facts (OMF), other than pure coincidence
On the other hand, on theism there is a plausible story according to which our moral intuitions correspond to OMF: God set up our evolution specifically this way to make our intuitions to correspond to moral facts because having moral intuitions corresponding to OMF is good and it's therefore represented in our moral intuitions.
I think I agree with that. Now, if probability of this story was itself higher than pure coincidence, that would be a point in favor of theism in regards of probability of objective morality.
But this doesn't appear to be the case! There is no particular reason why this story is more likely to be true compared to all the other possible stories. In other words, even though on theism we can reduce the improbability of us being right about OMF regarding boiling babies to the improbability of us being right about OMF regarding having correct moral beliefs, the latter is just as improbable as the former so no reduction in improbability happens.
> Just like the mere possibility of external world deception isn’t a positive reason to think our physical seemings are false or unreliable.
It is, though. It's prudent to consider this possibility and I think that initially some sort of subjective idealism is the correct ontological stance, given our epistemological situation. But then, as it's mentioned in the post:
> The thing about the external world hypothesis is that it compresses a lot of highly specific predictions (that turn out true) into a much simpler theory than alternatives. With moral realism I just don’t see why the moral facts being the way we think they are predicts our moral intuitions much better than deviant facts, given theism.
That’s not my claim. I’m thinking about it in terms of *defeaters*. I am not making the claim that I have some intuition-independent reason to think the theistic story on which we have moral knowledge is more probable than the deceptive stories. What I’m claiming is that our moral intuitions are defeated by naturalism, but not so by theism.
We start with moral intuitions about moral facts and the question is whether they have been defeated or not by something. On naturalism, we have a defeater. On theism, we do not. You’re right that we have no “3rd person” reason to think any of the theistic stories is more likely than the other. But it is mistaken to take this to be a defeater of your seeming that some moral fact is true — indistinguishable possibilities aren’t defeaters (and I bolster this by pointing out the analog of this point in the external world skepticism case). You respond by saying that it *is* a defeater in the external world context, and it’s just that the skeptical hypothesis is itself intrinsically less likely (because it is less parsimonious). But this strikes me as a clearly mistaken response in the context of external world skepticism, for the usual reason mentioned in the literature — the demon can simply be feeding you the seeming that parsimony is relevant for probabilistic determinations. You might respond that we must start with *some* principles to evaluate hypotheses, but the moment we do so, you’ve granted that indistinguishability based on one’s current evidence doesn’t constitute a defeater. Additionally, it’s just not at all clear to me that all skeptical hypotheses really are intrinsically less likely (they can posit significantly less entities, etc.)
I noticed that you are talking in terms of "defeaters", but I don't see how this term is useful. When I'm trying to translate your reasoning to the terms of conditional probabilities it really seems that it's not. Whether we have a "defeater" or not, probability of objective morality is still the same - as low as a simple coincidence. I'm not saying that in your terms we have a "defeater" of objective morality on theism, I'm saying that regardless of it it our reasoning about probabilities is the same.
So am I wrong here? Or is this "defeater" some completely different thing that does not affect probabilities in this case and our disagreement is that you believe that one should be reasoning about epistemology not just in terms of probabilities but also using this extra entity as well?
> it’s just that the skeptical hypothesis is itself intrinsically less likely (because it is less parsimonious)
A priori we can start from considering both hypothesises equally likely. But then as we collect more evidence about our reality, naturalism gets boosted by the fact that our world seem to be working on strict laws. There is no particular reason why would a demon make my perception of the world ordered like this, but there are all the reasons for it on naturalism.
We can update our deciever demon hypothesis to accomodate the evidence, saying that the demon just so happens to want the world to appear ordered, but this makes the hypothesis more specific and therefore less probable - only some part of probability mass of all possible deciever demons lies withing this more specific hypothesis.
> But this strikes me as a clearly mistaken response in the context of external world skepticism, for the usual reason mentioned in the literature — the demon can simply be feeding you the seeming that parsimony is relevant for probabilistic determinations.
It can, but whether a particular epistemical stance systematically produces truth is an empirical question. Try thinking that all things being equal the more complicated hypothesises is more likely to be true and see how well it works for you.
> Additionally, it’s just not at all clear to me that all skeptical hypotheses really are intrinsically less likely (they can posit significantly less entities, etc.)
Care to bet in advance which parts of the world are compressed like that? So far it really seems that the universe is bruteforcing everything and not saving any computations at all.
The important point here is that we can actually make accurate predictions about the behaviour of reality. We consult our naturalistic model and it systematically works very well.
To match this, we have to postulate that the demon just outright created the perfectly coherent simulation, including which epistemical stances are rational, that works exactly as if it was a naturalistic universe. And at this point our sceptical hypothisis becomes indistinguishable from naturalistic hypothesis + a demon did it. Which is just fine? The demon part is weird and doesn't fit our usual epistemology, but hey we can register it as a bold prediction, that our epistemic techniques stop working at this particular instance, which itself is very unlikely to the best of our knowledge, but still possible. Essentially it all added back up to normality.
In reverse order:
>”Try thinking that all things being equal the more complicated hypothesises is more likely to be true and see how well it works for you”
This doesn’t seem to engage with the point I made — that our taking this stance wrt hypotheses seems to consistently work out can be explained perfectly by the posited scenario, namely, that the demon feeds you this seeming. You need to simply start with some principles about how to evaluate hypotheses.
>”I'm not saying that in your terms we have a "defeater" of objective morality on theism, I'm saying that regardless of it it our reasoning about probabilities is the same”
The structure of defeaters is really crucial to the point I am making. The reason is because I think we all *need* to have the following picture: there are some beliefs we have that are justified (and thus if true basically constitute knowledge) in virtue of the fact that we have a seeming in their favor that is not subject to any defeaters. Now, it’s a tricky question whether this commitment to defeaters amounts to a difference in a Bayesian analysis — that’s something debated among epistemologists and I don’t exactly know where I land. But what I’ve always thought is that it doesn’t matter, at least for purposes like ours here.
The reason is this: if the structure of defeaters does make a difference to the conditional probabilities of hypotheses, it’s best understood as some sort of analysis of prior probabilities. If not, the relevant point is the conditional (if seems that X & no defeaters —> justified in believing X) still needs to be true for one to be justified in, e.g., believing in the external world in light of skeptical scenarios. As long as the conditional is *true*, I can make the point I made in my original comment. Other approaches to skepticism inevitably need to pull on a principle like this in order to justify the standards by which they’re evaluating hypotheses in the first place (this is the first point I made in this comment).