Responding to Schmid’s AMA on the Moral Argument Ranking

1. Introduction

One argument for theism, the argument from moral knowledge, uses moral knowledge as evidence for God’s existence. Another argument for theism, what’s sometimes called “the moral argument,” uses the existence of objective morality (roughly, the idea that there are moral truths that hold independently of human opinion) as evidence for God’s existence. Philosopher Joe Schmid ranked the argument from moral knowledge much higher than the moral argument. My opinion is that this ranking was flawed and in an AMA (ask me anything) Schmid responded to my question regarding this at around 5:30 to 22:58.

2. Philosophical Background

2.1 The Arguments

The argument from moral knowledge uses the existence of moral knowledge to argue for theism along the lines of this:

(1)	If God does not exist, then moral knowledge does not exist.
(2)	Moral knowledge does exist.
(3)	Therefore, God exists.

What is sometimes called “the moral argument” goes something like this:

(4)	If God does not exist, then objective morality does not exist.
(5)	Objective morality does exist.
(6)	Therefore, God exists.

The above moral argument is similar to one that Christian philosopher William Lane Craig (WLC) promoted. Note that (1) and (4) are conditional premises due to their if-then structure.

2.2 Moral Oughts and Moral Non-naturalism

The Humean ought is that type of ought that is distinct from the purely descriptive, in contrast to “oughts” that are identical to something that is purely descriptive. An example of a “descriptive ought” is when “If you want to do well in school, you ought to study” just means “As a matter of practical necessity, you need to study to do well in school,” where the “ought” in this context expresses that some possible action (studying) is practically necessary to achieve some goal (doing well in school). An example of a Humean ought, i.e., that type of ought that does not have only descriptive qualities, might be someone saying “You shouldn’t torture puppies.” The term “Humean ought” comes from David Hume’s famous is/ought problem where he noted the apparent difficulty of deducing an “ought” from an “is,” and where he called the “ought” an entirely different relation from the “is” in the final paragraph of book 3, part 1, section 1 of A Treatise of Human Nature. Notably, Hume specifically had the moral ought in mind for his is/ought distinction, and the type of “ought” used in morality is the Humean ought (at least, for ordinary moral semantics). Deontic morality is morality having to do with obligation; e.g., moral wrongness (since an act is morally wrong for someone only if they morally ought not do it), and deontic morality uses the Humean ought.

A notable implication of moral oughts being Humean oughts is that, as the Folk Moral Semantics Argument explains, this leads to morality being non-natural (i.e., moral properties like moral wrongness do not exist solely as part of the natural, physical world). We can use this fact to help justify the argument from moral knowledge’s first premise and the moral argument’s first premise.

2.3 Justifying the Argument from Moral Knowledge

In justifying the first premises of both arguments, we can use a theorem of mathematics and propositional logic that says “Given A, probably C” entails “Probably, if A then C.” For example, if moral knowledge is unlikely given God’s nonexistence, then premise (1) is probably true; and if objective morality is unlikely given God’s nonexistence, then premise (4) is probably true. In this context, the type of probability I’m using here is epistemic probability, which is the same sort of probability we use when we say, “The evidence suggests that the accused is probably guilty,” i.e., the probability that something is the case given some state of information.

Why think moral knowledge is unlikely on atheism? First note how odd it is that we even know that morality exists. Morality is non-natural and empirically undetectable. To illustrate what I mean by morality being empirically undetectable, imagine a moral nihilist (who disbelieves in moral wrongness) and a moral objectivist both observe some jerk kicking a dog just for fun. Both agree on all the same psychological and physiological facts (e.g., the dog felt pain and suffered minor injury). The moral nihilist says “I don’t think moral wrongness is attached to that action,” whereas the moral objectivist says “I think moral wrongness is attached to that action.” There is no empirical way to determine who is right here; both sides agree on all the same empirical facts. In this sense, moral wrongness is empirically undetetable.

So how do we know that moral wrongness exists? In practice we rely on moral intuition, but since moral wrongness is non-natural, barring supernatural intervention (e.g., God) our brains would deliver the same intuitions of moral wrongness existing regardless of whether morality existed, and this would seem to undercut such intuition from properly justifying our belief in morality. Given naturalism (the view that he supernatural does not exist) moral knowledge is infeasible. The atheist could get around this problem by ditching naturalism and saying we humans have supernatural powers of clairvoyance to detect these invisible, non-natural properties, but that seems far-fetched. Given atheism, it is unlikely that we have moral knowledge, in which case premise (1) is at least probably true.

2.4 Justifying the Moral Argument

The existence of objective morality is to be expected if God exists; e.g., God is objectively morally good. But it seems to me that moral objectivism is unlikely on atheism. Why? Against moral objectivism, the atheist could use what’s called the argument from queerness that atheist philosopher J.L. Mackie promoted. Quoting Mackie in chapter 1 section 9 of his book Ethics: Inventing Right and Wrong where speaks of prescriptive objective values:

If there were objective values, then they would be entities or qualities or relations of a very strange sort, utterly different from anything else in the universe. Correspondingly, if we were aware of them, it would have to be by some special faculty of moral perception or intuition, utterly different from our ordinary ways of knowing everything else.

To get a general idea of what “queerness” means in this context, imagine someone telling you that space aliens landed in a flying saucer somewhere in New Jersey. Such an incident is “queer” in the sense that it wildly differs from the types of things we know exist such that we give it a low prima facie probability (i.e., a low probability in the absence of any evidence to the contrary). Its queerness is such that if your next-door neighbor told you this happened, you would probably still disbelieve it, whereas you might believe your next-door neighbor if they made a mundane claim like having coffee this morning. Similarly, we would give a low prima facie probability to someone’s claim that an invisible, incorporeal dragon is floating in their garage because such an incident is queer in the relevant sense.

Mackie’s argument from queerness has two components: epistemological (roughly, having to do with knowledge) and metaphysical (roughly, having to do with what exists). Moral qualities like moral wrongness are epistemologically queer on atheism; we’d need something like supernatural clairvoyance to know they exist, which of course is utterly different from how we know anything else. Moral qualities are also metaphysically queer on atheism; they’re invisible, nonphysical, empirically undetectable, and (as mentioned earlier) we’d need something like supernatural clairvoyance to know they exist. On atheism at least, this is wildly different from the types of things we know exist, akin to an invisible, nonphysical, empirically undetectable dragon.

How I would argue for premise (4) is this: combine the argument from moral knowledge (given atheism, we don’t have moral knowledge) with Mackie’s argument from queerness (given atheism, morality is “queer” in a way to make it prima facie unlikely) to argue that given God’s nonexistence, objective morality probably doesn’t exist, which would make premise (4) at least probably true.

3. Question And Response

3.1 The Question

My question to Schmid was this:

You rated the argument from moral knowledge highly but a Craig-style deductive moral argument remarkably low despite WLC's report on its dialectical effectiveness, and you implied the moral argument's conditional premise was weak. By my lights the argument from moral knowledge isn’t that far above the moral argument, because what about combining the argument from moral knowledge with Mackie's queerness objection (moral objectivism would be queer on atheism) to argue that objective morality is epistemically unlikely given atheism? This would justify the conditional premise [being] probable, and I can’t think of a great atheist response to this justification conditional premise that wouldn't involve attacking the argument from moral knowledge in some way. After all, if it’s true that moral properties are non-natural (if we use the Humean ought for our moral semantics), without God it seems we’d need something like supernatural clairvoyance to detect these invisible, non-natural moral properties, and that seems queer enough on atheism to be far-fetched and unlikely.

As mentioned earlier, Schmid responded to my question regarding this at around 5:30 to 22:58.

3.2 Four Possible Atheist Responses to Metaphysical Queerness

Going through four possible atheist responses he brought up and adding my rebuttals:

1. The atheist can go for metaethical naturalism.
   • Moral naturalism is the idea that moral facts and properties are natural facts and properties. Yet moral naturalism isn’t really an option if we use the Humean ought for our moral semantics, which leads to moral naturalism being false. Apparently, I didn’t make this sufficiently clear in my question.
2. Why think that objective non-natural facts are queer under atheism?
   • This was kind of already answered in my question when I said “without God it seems we’d need something like supernatural clairvoyance to detect these invisible, non-natural moral properties, and that seems queer enough on atheism to be far-fetched and unlikely.” This doesn’t generally hold true for epistemic facts (facts having to do with knowledge, justification, rationality, etc.), mathematical facts, and so forth (i.e., we don't need something like supernatural clairvoyance to know that stuff).
3. Moral facts are equally queer on theism.
   • Objective morality is not at all queer given certain standard conceptions of God; e.g., God being objectively morally good. Moreover, moral intuition delivering moral knowledge is not at all queer on Christian theism (see Romans 2:14-15).
4. Why think morality being queer on atheism makes it unlikely on atheism?
   • This was already answered in section 2.4 of this article, but to recap: something is queer in the relevant sense if it is wildly different from the types of things we know exist such that it is prima facie unlikely. This is just how we normally operate (e.g., for space aliens landing in New Jersey and invisible incorporeal dragons floating in people’s garages), and we seem to be rational in doing so.

I thus think the responses are woefully inadequate.

3.3 A Priori Moral Knowledge?

Schmid also says this (at least in a paper onscreen) at around 14:55 to 15:49, and where a priori means “knowable not on the basis of experience” (e.g., the fact that two plus two equals four is knowable by seeing the relations of ideas, as opposed to needing to believe it on the evidential basis of empirical experience):

In fact, if objective moral facts are a priori knowable (as non-naturalists would surely accept), then they are not improbable under atheism, since any epistemic probability function should assign probability 1 to all a priori truths.

As I said, a priori refers to “independent of experience,” which is in contrast to a posteriori knowledge which is knowledge that is believed on the basis of experience (e.g., the way one might believe that it is now raining outside). However, not everyone agrees on what should and should not count as “experience” in this context, though it’s uncontroversial that sense experience would be included. For example, the fact that Sam knows a computer monitor is in front of him is not knowable a priori since Sam relies on his sense experience to know this. Whereas “experience” would not include the way we mentally “see” that certain things are true by definition, such as the way we mentally “see” that all bachelors are unmarried after knowing the definitions of words like “bachelor” and “unmarried.” However, memory and introspection (e.g., Sally’s mental perception that she is thinking of the number seven) would seem to count as experience.

Some beliefs are not believed on the basis of experience but whose probability is not 1. For example, consider our intuition that the future will relevantly resemble the past, something that we can call the principle of induction. Justifying induction is a famous problem in philosophy. You might think we can justify it on the basis of experience, saying like “The principle of induction (roughly, that something will probably hold in the future if it held in the past) will probably hold for the future because this principle held in the past,” but this argument uses the principle of induction to justify the principle of induction, which is circular reasoning. In practice at least, we accept the principle of induction on the basis of intuition; the principle just seems true to us.

Where does moral experience (that is, moral intuition) fall into this? Well, that’s tricky. David Hume seemed to think that the moral sense counts as some kind of experience. While the existence of objective moral truths is surely rational to believe despite not being empirically undetectable, it’s not exactly true by definition the way All bachelors are unmarried is. Moral intuition is more akin to the intuition that our memory are at least sometimes reliable or the intuition that the future will probably relevantly resemble the past.

Like many of us, philosopher G.E. Moore held that we accepted moral truths on the basis of intuition in the sense of an intuitive seeming-to-be-true experience. To quote from the Stanford Encyclopedia of Philosophy:

Moore held that what justifies one in believing a fundamental moral principle is that the principle is self-evident [in the sense that it seems evident to the self without being inferred on the basis of other beliefs] and that it appears to be true, i.e., that one has an intuition of it. But this intuition, as an appearance of truth, is clearly a kind of experience, so it seems that one’s justification for an intuitive belief in a fundamental moral principle will not be independent of experience. Hence, to regard this justification as a priori, one must make an exception for this specific kind of experience. While Moore famously denied hedonism, this was because he denied that pleasure is the only good. He accepted that pleasure is good, and this proposition nicely illustrates the tension inherent in the modified standard view of a priori justification, and indeed the standard view, as well as Moore’s a priori moral epistemology. Presumably one must experience pleasure to acquire the concept of pleasure. But then how exactly is it that one is justified in believing that pleasure is good? Is it that once one has the concepts of pleasure and goodness, one believes that pleasure is good simply on the basis of understanding this proposition? Or is it that once one understands this proposition, one has a rational intuition of its truth? Or is it instead that when one experiences pleasure one also experiences its goodness? If something like the latter, the justification for our belief that pleasure is good seems more empirical than a priori.

My first instinct is to categorize moral experience as the sort of experience in the a posteriori category (though I suspect this is not a popular view among philosophers). Regardless, even if we categorize moral experience as not counting as the right sort of experience to be a posteriori, I don’t think we should assign the probability to our moral beliefs to be 1 for the same reason we shouldn’t for the principle of induction and the belief that our memory are at least sometimes reliable; it just doesn’t seem to be the sort of thing that is absolutely certain in the way that All bachelors are unmarried or One plus one equals two is. Don’t get me wrong, I think it is very likely that my intuition of there being some kind of supremely authoritative non-natural reality that imposes the moral prohibition to not torture infants just for fun is very likely true—just as I think it is very likely that my intuition of my memory being at least somewhat reliable is very likely true. It’s just that the high probability isn’t quite equal to 1 in the same way that All bachelors are unmarried is.

3.4 Possible Atheist Responses to Epistemic Queerness

Some philosophical background before going forward. A possible world is a complete description of the way the world is or could have been like. For some proposition p that is possibly true (and thus is true in some possible world), one could think of “belief safety” as “In all possible worlds appropriate similar to the actual world, if I believe p, then p is true” (very roughly, belief safety means if I believe it’s true then I couldn’t have easily been wrong). For some p that is possibly true, one could think of “belief sensitivity” as “In all possible worlds appropriate similar to the actual one, if p is false then I would not believe that p” (very roughly, belief sensitivity means if the belief were false I couldn’t have easily believed that the belief is true). Some people belief that various moral truths are metaphysically necessary in the sense that they hold in all possible worlds. For moral truths that are metaphysically necessary, all such truths would be “safe” beliefs (since in all possible worlds in which they are believed, they are true, due to them being true in all possible worlds).

Schmid notes that when it comes to the epitemological component to the argument from queerness, the atheist can give these responses.

1. Not accept that moral properties are non-natural.
   • As already explained, the question has in mind Humean ought moral semantics, in which case moral naturalism is false.
2. A queer mental faculty (like supernatural clairvoyance) is not needed for moral knowledge. One might think that belief sensitivity and/or belief safety is all we really need.
   • To illustrate why this doesn’t work, suppose there exists a metaphysically necessary, incorporeal, causally inert dragon named Carl. By chance, a species that we can refer to as “Believers” has evolved the intuition that Carl exists, though because Carl is causally disconnected from the universe, the Believers would have the intuition that Carl exists regardless of whether Carl had in fact existed (Carl is metaphysically necessary and could not have failed to exist, but if he had not existed, his absence would have zero effect on whether the Believers had their dragon intuitions). Even if the Believers’ belief in Carl is sensitive and safe, their dragon belief still wouldn’t count as knowledge. The same, I think is true for moral beliefs if naturalism is true. We need the supernatural to have moral knowledge. Moreover, there does not appear to be a relevant difference between the “Dragon Scenario” and what we could call the “Naturalism Scenario” (where naturalism and moral objectivism are both true) in which the Believers do not have dragon knowledge but moral objectivists have moral knowledge.

Schmid makes a great point though that I hadn’t considered. It seemed to me that moral knowledge on atheism is epistemically unlikely because it would require a queer way of knowing (basically, supernatural clairvoyance), but there’s another approach: one could deny that the relevant modal conditions (belief sensitivity or belief safety) would likely obtain if naturalism is true. Belief sensitivity and belief safety aren’t “queer” relations for beliefs. So one wouldn’t necessarily have to argue from queer epistemology to support moral knowledge being unlikely on atheism. I concede Schmid’s point, but only to a certain extent, because it really does seem to me that some kind of queer epistemology is needed as illustrated in the Dragon Scenario. Though Schmid does agree with me that the atheist would need some kind of “sui generis faculty by which we can grasp moral facts directly” and that this is very “spooky” under naturalistic atheism. So Schmid’s views and mine are at least somewhat similar.

4. Conclusion

I’m grateful that Schmid took the time to answer my question in his AMA. However, I think the four responses the atheist could make for metaphysical queerness are woefully inadequate, and in some cases seem to rely on not quite accurately understanding the question in the first place. Even if moral knowledge via experience can be correctly categorized as a priori, it is not the kind of a priori knowledge that is absolutely certain in the way that All bachelors are unmarried is. The described possible atheist responses for morality being epistemically queer seem to fail. Still, it’s worth noting that Schmid seems to agree with me that there is a kind of epistemic queerness for moral knowledge on naturalistic atheism.