I mean, it seems that you grant that, regarding priors, there is a similar argument along basically the same lines. I agree that calling it a moral knowledge argument is not productive as it's more broadly about a priori knowledge.
I don't see anything wrong with the following inference:
1) I know P
2) were Q true I wouldn't know P
3) So Q is false.
In the case of moral knowledge, it seems that we can have theory-neutral knowledge that we have moral knowledge. If a view comes along that undercuts that moral knowledge, well, we know it's false! We don't need to say "well, if that view were true you'd think you had moral knowledge," because we know that we do have moral knowledge of a sort we wouldn't if that view were true.
I do think there's something wrong with the inference, if (2) is true because Q undercuts your justification for P. I think this is made blatantly obvious by the watch-example I give. The person there could make the same inference:
1) I know it's 12:59
2) If my watch didn't have batteries, I wouldn't know it's 12:59
3) So my watch have batteries.
The reason this inference is wrong is that your evidence isn't that you *know* that it's 12:59, but that you have some evidence (the hands of the clock are in a certain position) that would give you knowledge *given* certain background assumptions (that there's a battery in the clock, and it's properly set).
Likewise, with atheism and moral knowledge, your evidence isn't that you know moral facts, but that you have evidence that would give you knowledge of moral facts given theism. But that means that you can't use the evidence in question to infer from atheism to theism--the evidence is predicted on both theories, it's just that it doesn't justify the same things on both theories.
The inference *would* be fine if the reason you wouldn't know P given Q was that you wouldn't have the same evidence that you have now--for example, if P is "Jones' fingerprints are on the Gun" and Q is "Jones' didn't commit the murder". But here it's because you wouldn't even have the evidence indicating P (your visual experience of the fingerprints on the gun) if Q is true.
The only way the inference would be generally fine, would be if it was formulated:
1) I have evidence E
2) were Q true, I wouldn't have E
3) So Q is false
But in this case it doesn't work for moral knowledge, as atheism predicts the evidence.
The difference is that in the watch example, your belief in the time is not a basic belief that's intrinsically justified, but only derivatively justified from the general reliability of watches. You can know that you know P without first knowing the process by which you know P.
Well, consider another example then: your belief that you have hands. This is presumably a sort of Moorean basic belief; you don't need to know how eyes, brains, etc., work in order to be justified in believing it.
But imagine that you lost your arms in a tragic sword-fighting accident. You became so miserable from this, that you decided to plug yourself into an experience machine, where you would be appeared to as if you have arms (including hands), and you now remember having done this. Here you lose your justification for belief in having hands, and you shouldn't Moorean shift to thinking that you're actually at base-level reality.
But this looks relevantly analogous to the moral knowledge argument: You are appeared to as if it's wrong to pull the tails of cats. But now you realize that evolution+atheism (which you judge to be highly probable) entails that this appearance has no connection to whether it's in fact wrong. In that case you should give up the belief, rather than Moorean shift away from evolution or atheism.
Oh, well then you can just imagine that you become super convinced of arguments for idealism, or that God telepathically tells you that you're a BIV. Now you're no longer justified in believing in the external world. Again, I don't see any relevant difference from the moral knowledge argument.
I think you’re missing part of the argument. Or, at least, there’s some version of this argument that is more charitable than the one you presented. Here’s my best shot at reconstructing it (given a few minutes lol):
The idea here starts from the fact that a deep intuition in some P is a really strong reason to believe that P is in fact true (perhaps unless you have some defeater). Given that we think morality is like P in this scenario, we can say something like: we know with high probability that moral facts are true.
Now we have some evidence and can applies Bayes Theorem (or something similar). Given that we know M (whereas M is the proposition that there are moral facts), this might have implications on our other beliefs like god.
Moral facts being true conditional on atheism seems quite low. Moral facts being true conditional on theism seems quite high. Therefore, we should update our beliefs in a positive direction on god upon learning that moral facts are real.
The problem is that you exactly *do* have a defeater for your intuition--namely that you find out that whether you have the intuition or not, doesn't track the content of the intuition (given your other beliefs). So now that intuition no longer gives you strong justification for M.
This doesn’t seem entirely true. Sensitivity, first of all, is (albeit arguably) only a sufficient condition for believing true things with respect to morality. Secondly, it’s not clear that they don’t track truth - similar to math, under the realist model, it may be tough/ impossible to imagine worlds in which it wasn’t the case that the moral facts were true. Maybe this can be resolved if you give a little more information about what you mean by this defeater (for example, what do you mean by “other beliefs” here?).
All this might be right, but the person forwarding the argument from moral knowledge shouldn't grant this. This is because the argument relies on a debunking argument working *given* atheism. So they would want to say that if atheism were true, you would have a defeater for your belief. But then the problem is simply identifying what reason we then have to think that we *don't* have this defeater. And as far as I can tell, the only thing you can do is insisting that it's obvious that you don't have a defeater. The problem is that this would be equally obvious, regardless of whether you *do* have a defeater or not.
Great post. I've always had the same issue with epistemological arguments for God. They argue that some sort of skepticism is true given atheism, but they don't actually give you evidence against skepticism (on the contrary, if the argument is successful, it's strong evidence for skepticism). So it doesn't make sense to update your beliefs away from atheism. But you've described it in a clearer manner than I probably could.
Suppose we grant phenomenal conservatism. If you were convinced most people had very different intuitions/seemings than you, and that most people did not, pretheoretically or on reflection, report that it seemed to them that things were moral or immoral in a distinctively realist way, would this undermine your confidence in moral realism?
I suppose I'm wondering how much it personally seeming to ourselves that something is true is serving as evidence for the view in question, and how much one's confidence in a position is driven by the belief that it seems the same or a similar way to others.
Ok, but if moral realism is false then we have no good reason to believe that moral realism is false.
The idea that moral realism is true makes me happy, and I think it makes most people happy. Why should I care if it's wrong? Because I "ought" to care about truth?
Well, *theism* is a defeater-defeater. But the move of inferring from atheism not being able to support moral knowledge to theism I think is a Moorean shift: you are presented with an argument of the form:
I mean, it seems that you grant that, regarding priors, there is a similar argument along basically the same lines. I agree that calling it a moral knowledge argument is not productive as it's more broadly about a priori knowledge.
I don't see anything wrong with the following inference:
1) I know P
2) were Q true I wouldn't know P
3) So Q is false.
In the case of moral knowledge, it seems that we can have theory-neutral knowledge that we have moral knowledge. If a view comes along that undercuts that moral knowledge, well, we know it's false! We don't need to say "well, if that view were true you'd think you had moral knowledge," because we know that we do have moral knowledge of a sort we wouldn't if that view were true.
I do think there's something wrong with the inference, if (2) is true because Q undercuts your justification for P. I think this is made blatantly obvious by the watch-example I give. The person there could make the same inference:
1) I know it's 12:59
2) If my watch didn't have batteries, I wouldn't know it's 12:59
3) So my watch have batteries.
The reason this inference is wrong is that your evidence isn't that you *know* that it's 12:59, but that you have some evidence (the hands of the clock are in a certain position) that would give you knowledge *given* certain background assumptions (that there's a battery in the clock, and it's properly set).
Likewise, with atheism and moral knowledge, your evidence isn't that you know moral facts, but that you have evidence that would give you knowledge of moral facts given theism. But that means that you can't use the evidence in question to infer from atheism to theism--the evidence is predicted on both theories, it's just that it doesn't justify the same things on both theories.
The inference *would* be fine if the reason you wouldn't know P given Q was that you wouldn't have the same evidence that you have now--for example, if P is "Jones' fingerprints are on the Gun" and Q is "Jones' didn't commit the murder". But here it's because you wouldn't even have the evidence indicating P (your visual experience of the fingerprints on the gun) if Q is true.
The only way the inference would be generally fine, would be if it was formulated:
1) I have evidence E
2) were Q true, I wouldn't have E
3) So Q is false
But in this case it doesn't work for moral knowledge, as atheism predicts the evidence.
The difference is that in the watch example, your belief in the time is not a basic belief that's intrinsically justified, but only derivatively justified from the general reliability of watches. You can know that you know P without first knowing the process by which you know P.
Well, consider another example then: your belief that you have hands. This is presumably a sort of Moorean basic belief; you don't need to know how eyes, brains, etc., work in order to be justified in believing it.
But imagine that you lost your arms in a tragic sword-fighting accident. You became so miserable from this, that you decided to plug yourself into an experience machine, where you would be appeared to as if you have arms (including hands), and you now remember having done this. Here you lose your justification for belief in having hands, and you shouldn't Moorean shift to thinking that you're actually at base-level reality.
But this looks relevantly analogous to the moral knowledge argument: You are appeared to as if it's wrong to pull the tails of cats. But now you realize that evolution+atheism (which you judge to be highly probable) entails that this appearance has no connection to whether it's in fact wrong. In that case you should give up the belief, rather than Moorean shift away from evolution or atheism.
You have a Moorean belief (maybe) in there being an external world but not in you, in fact, having hands.
Oh, well then you can just imagine that you become super convinced of arguments for idealism, or that God telepathically tells you that you're a BIV. Now you're no longer justified in believing in the external world. Again, I don't see any relevant difference from the moral knowledge argument.
I think that if you consider the external world a Moorean fact then you shouldn't become convinced of cases for you being a BIV or idealism.
I think you’re missing part of the argument. Or, at least, there’s some version of this argument that is more charitable than the one you presented. Here’s my best shot at reconstructing it (given a few minutes lol):
The idea here starts from the fact that a deep intuition in some P is a really strong reason to believe that P is in fact true (perhaps unless you have some defeater). Given that we think morality is like P in this scenario, we can say something like: we know with high probability that moral facts are true.
Now we have some evidence and can applies Bayes Theorem (or something similar). Given that we know M (whereas M is the proposition that there are moral facts), this might have implications on our other beliefs like god.
Moral facts being true conditional on atheism seems quite low. Moral facts being true conditional on theism seems quite high. Therefore, we should update our beliefs in a positive direction on god upon learning that moral facts are real.
The problem is that you exactly *do* have a defeater for your intuition--namely that you find out that whether you have the intuition or not, doesn't track the content of the intuition (given your other beliefs). So now that intuition no longer gives you strong justification for M.
This doesn’t seem entirely true. Sensitivity, first of all, is (albeit arguably) only a sufficient condition for believing true things with respect to morality. Secondly, it’s not clear that they don’t track truth - similar to math, under the realist model, it may be tough/ impossible to imagine worlds in which it wasn’t the case that the moral facts were true. Maybe this can be resolved if you give a little more information about what you mean by this defeater (for example, what do you mean by “other beliefs” here?).
All this might be right, but the person forwarding the argument from moral knowledge shouldn't grant this. This is because the argument relies on a debunking argument working *given* atheism. So they would want to say that if atheism were true, you would have a defeater for your belief. But then the problem is simply identifying what reason we then have to think that we *don't* have this defeater. And as far as I can tell, the only thing you can do is insisting that it's obvious that you don't have a defeater. The problem is that this would be equally obvious, regardless of whether you *do* have a defeater or not.
(Btw, I’m a atheist, don’t like phenomenal conservatism, and a moral anti realist. Just thought that this was a good clarification).
Great post. I've always had the same issue with epistemological arguments for God. They argue that some sort of skepticism is true given atheism, but they don't actually give you evidence against skepticism (on the contrary, if the argument is successful, it's strong evidence for skepticism). So it doesn't make sense to update your beliefs away from atheism. But you've described it in a clearer manner than I probably could.
Thank you, glad you enjoyed it :)
Suppose we grant phenomenal conservatism. If you were convinced most people had very different intuitions/seemings than you, and that most people did not, pretheoretically or on reflection, report that it seemed to them that things were moral or immoral in a distinctively realist way, would this undermine your confidence in moral realism?
I suppose I'm wondering how much it personally seeming to ourselves that something is true is serving as evidence for the view in question, and how much one's confidence in a position is driven by the belief that it seems the same or a similar way to others.
Ok, but if moral realism is false then we have no good reason to believe that moral realism is false.
The idea that moral realism is true makes me happy, and I think it makes most people happy. Why should I care if it's wrong? Because I "ought" to care about truth?
Calling defeater-defeaters Moorean shifts hurt my soul
Well, *theism* is a defeater-defeater. But the move of inferring from atheism not being able to support moral knowledge to theism I think is a Moorean shift: you are presented with an argument of the form:
1. P
2. If P then not-Q
3. Therefore not-Q
And then insist:
4. If P then not-Q
5. Q
6. Therefore not-P