Let's recap the argument:

1. It appears that we are designed to kill animals to eat their meat.

2. If it appears that we are designed to Φ, it is not always immoral to Φ.

3. (Conclusion, From 1 + 2) It is not always immoral to kill animals to eat their meat.

   

Why Believe Premise 2?

In my original article, I responded to criticisms of the premises, but didn't give much positive reason to believe Premise 2. As such, many critics said ‘Sure, you might have good responses to the objections, but why believe it in the first place?’ Here are some accounts of why you might believe Premise 2:

Theism

If you believe humans were designed by an all-good God, you will probably also consider it likely that their design is morally acceptable. This is not an ‘auxiliary hypothesis bolted onto’ theism. It is something we can expect to be true given theism.

Another way of putting this is that if moral agents were designed to go around torturing babies for fun, that would be pretty strong evidence that an all-good God had NOT designed those moral agents. Of course this alternative framing suggests that you can modus tollens my modus ponens in order to make an argument against the existence of God, but that would only work if you were already more confident about veganism than about God. I, personally, think there are stronger independent arguments for theism than there are independent arguments in favour of not-3. Therefore, I should stick to my modus ponens and conclude that KATEM is sometimes permissible.

Ad hominems

1There are two problems for this theistic reasoning:

Skeptical Theism

I am a skeptical theist. This means that I can’t use noseeum2 inferences about God’s reasons for action. Therefore, my theistic support for Premise 2 has to be based on some other kind of inference. It won’t do to say ‘I see a reason for God to do this and I see no outweighing reason not to’.

Luckily for me, I can use a non-noseeum inference, as I said last time. I can appeal to Genesis 9:3 where God says it is permissible to eat meat. This is an inference using God’s testimony about morality, rather than a noseeum inference. So, if your theism accepts Genesis 9:3 (if you are Christian, Muslim, Jewish, Bahá’í, or something else I haven’t heard of), then the skeptical theism worry need not bother you here. Furthermore, if you are any kind of theist that doesn’t accept skeptical theism to begin with (instead using theodicies to respond to the problem of evil) then this need not worry you either.

The Edenic undercut

In a beautifully written piece, Okoli Akeju issued a further problem for me. As a Christian, I accept Genesis, which describes how Adam and Eve did not eat meat in Eden. This is a case, Akeju says, where Premise 2 is not true. This is ‘the Edenic undercut’.

However, this fails to engage properly with Premise 2. Premise 2 says it isn’t always wrong to KATEM. Eden only constitutes one moment when it was wrong to KATEM, rather than a possible world where it was always wrong. Thus, Premise 2 is unaffected.

Here’s a thought, though: Eden was what God intended the world to always be like, and so there is a possible world where Adam and Eve didn’t sin, and therefore where it was always wrong to eat meat.

I don’t think this works either. My reason is that I think God did not intend the world to remain Edenic forever. Following Aquinas and Plantinga, among others, I think a theistically ideal timeline is one which includes the Fall, as it allows for various goods stemming from evil. Therefore, the timeline God planned for is this one3. Therefore, the Edenic undercut does not succeed. Theism can give reason to believe Premise 2.

The ‘Whatever’ view

The theistic view is my own. The ‘Whatever’ view is one which I can imagine others holding. In fact, if I understood him correctly, one commenter actually proposed it under my original post! The idea here is that if a creature has evolved to do something that we don’t like, we shouldn’t hold it against them and consider the action immoral - it’s just something they need to do to survive. We should just say ‘Whatever, live and let live’. This might be the view of somebody who isn’t too fussed by mosquitoes, for example.

I’m not 100% sure how to characterise this view to make it totally clear. One way I put it in conversation with somebody was that it tugs on the thought that for something to be immoral it must be something you can change about yourself. Mosquitoes can’t change the fact that they are designed to drink our blood, so we can’t say they’re wrong to do it.

In my original post, I considered the ‘Aliens’ objection to Premise 2 - there could be aliens that have evolved to do nasty things to humans and this could tell against 2. The ‘Whatever’ view deals nicely with this objection by just saying that the aliens aren’t acting immorally by trying to carry out their life cycle. The view would probably also license us killing the aliens to stop them.

In the aforementioned Akeju piece, Akeju pointed out that biologists in a lab could alter DNA so that people were designed to do bad things. The ‘Whatever’ view would also say that those genetically modified people were not acting wrongly in carrying out their design. This is not a bug in the view but a feature.

Evolutionary morality

If you believe morality is a product of evolution, then you might also think that Premise 2 is true. The idea here is that that which is evolutionarily advantageous is also morally permissible. What is good for you in an evolutionary sense is what determines what is moral. Another view that might support 2 which you could group in here is virtue ethics - what you need to do to flourish is what is permissible.

Does evolutionary morality also suffer from the ‘aliens’ objection? Some people will be happy to accept that evolutionary morality does let the aliens do their thing, thinking in a similar vein to the ‘Whatever’ attitude. Other evolutionary moralists might think that evolution only applies to humans as a human institution, and so it does not give the aliens a warrant to be nasty.

Share

Who does the argument convince?

One commenter worried that the argument only convinces Biblical literalists.

God could allow evil for various reasons. Genesis 9:3 could be meant allegorically or is not applicable to modern humans anymore. Or I could be a non-Christian, deist, or atheist. I kept waiting to see any argument addressing those groups, but unfortunately it never came.

Biblical literalism means that you think the author’s original intent for the text is true. I don’t think you need to be a Biblical literalist to think that Genesis 9:3 supports eating meat: Suppose I wrote an allegory, in which I had a character, stated to be morally perfect and omniscient, declare that it is morally permissible to torture babies. If you told me off for promoting torture, it obviously wouldn’t do to say ‘I meant it allegorically!’

What if the permission is no longer applicable to modern humans? If that were the case, I would expect to see an explanation of this somewhere in the Bible to clear up ambiguity, since that is what happens with pretty much every other difference in Old and New Covenants in the Bible. I certainly wouldn’t expect to see Christ eating fish, or declaring that ‘that which goes in by the mouth does not defile’, or telling Peter to eat bacon!

As regards the commenter’s worry about the effect of the argument on non-Christians, deists, or atheists, see the above section for some non-theistic supports for Premise 2.

Which veganism?

Other commenters complained that my piece did not critique their form of veganism.

I specified exactly what form of veganism I was attacking in the subtitle of the article - that which thinks KATEM is always wrong. It would be way above my pay-grade to offer one master argument that defeated all forms of veganism. That’s because ‘veganism’ can mean a lot of things: it could refer to the belief that one should never KATEM, the belief that KATEM is always negative but not always impermissible, the belief that KATEM via factory farms is always wrong, the belief that KATEM via factory farms is always negative but not always impermissible, the belief that KATEM via factory farms is usually wrong. I could go on. I targeted the first of these possible meanings. It is no skin off my nose that I didn’t address the other views simultaneously.

The Naturalist fallacy

Many people thought I was committing a fallacy.

(There were more…) Luckily for me, I don’t care if you can stick a label on my argument and thereby call it a fallacy4! Since my argument is logically valid, you need to actually tell me why my premises are false in order to defeat it. #SorryNotSorry.

The Context

If you know what skeptical theism is, you can skip this section.

Many atheists propose the problem of evil as their strongest argument against perfect-being theism. The version most atheists endorse runs like this:

1. If God existed, there would be no gratuitous evil - that is, there would be no evil that was not justified in some way or necessarily connected to some greater good/prevention of worse evil that outweighed it.

2. There is in fact gratuitous evil.

3. Therefore, God does not exist.

While many theists propose theodicies which aim to explain why evils that appear to be gratuitous are not in fact gratuious, making the second premise false, there is an alternate way to argue against that premise: point out that God might have reasons for allowing evils which we shouldn’t expect to understand. This is meant to block most inferences that would allow you to form a warranted belief in the second premise2. This way of defending theism from the problem of evil is known as skeptical theism.

There are several arguments against skeptical theism. Let’s deal with them in the order in which they came.

Theoretical Defeat (Tettü’s Argument)

Tettü holds that

A good theory earns credibility by reducing mystery and increasing understanding.

Further,

Classical theism claims to be [explanatorily] fruitful[.]

Skeptical theism, however, Tettü argues, does not reduce mystery and increase understanding. Instead, it says

many features of the world will remain opaque because God’s reasons exceed our comprehension.

Here we can see, then, why Tettü thinks that skeptical theism admits defeat - a theory is meant to shed light on things. Instead, skeptical theism shrouds the world in darkness:

A theory that predicts in advance that vast domains of data will be inexplicable is simply a poor theory.

To summarise, Tettü’s view is that skeptical theism has no explanatory power because it doesn’t help us to understand things; it instead predicts that we wouldn’t understand things. In a similar vein, The Sacred & The Secular (hereafter TSTS) claims that

[…] skeptical theism is basically what happens [when] the apologist gives up but doesn’t want to say they gave up.

insinuating that skeptical theism does no theoretical work at all.

In response, I would contest Tettü’s account of what makes a theory credible. To see why, imagine two toddlers, Mikey and Alexa, arguing:

Mikey: Mummy and Daddy made me go to the dentist’s! It hurt! They must not love me, otherwise they wouldn’t have given me such an ouchie.

Alexa: I think they do love you!

Mikey: Then why did they do it?

Alexa: I don’t know…

Mikey: Then don’t say they love me if you don’t know why they did it. Nurr nurr you don’t know, you don’t know!

Alexa: But Mummy is much cleverer than me. Maybe she knows why!

Mikey: Well, unfortunately for you, Alexa, your hypothesis that Mummy and Daddy love me was meant to have explanatory power. So now that you have augmented that hypothesis with the additional assertion that Mummy knows more than me and so might have knowledge that I don’t, you have warped the theory such that it no longer serves its original purpose! It predicts that I wouldn’t know the reason for my dentist’s visit, when the whole point of the theory was to help me know things!

If I were Mikey’s dad, I would lecture him as follows:

No Mikey, the point of the ‘loving Mummy and Daddy’ theory is not to help you know things. The point of the theory is to give an accurate account of the world. It gains credibility by predicting things you already know to be the case. It doesn’t simply increase your understanding of other things; other things increase your understanding of the theory: your knowledge that the ice cream I bought you tastes good tells you that my love for you is strong enough for me to buy you nice ice cream; your knowledge that I gave you a shoulder ride tells you that I love you enough to hurt my shoulders for your pleasure.

The knowledge that I made you go to the dentist is not something which the theory immediately explains. But that doesn’t mean there is no possible explanation consonant with the theory.

When Alexa pointed out that Mummy is cleverer than you, she wasn’t negating the explanatory power of the theory. She was pointing out how the ‘intelligent Mummy and Daddy’ theory (a common auxiliary to the ‘loving Mummy and Daddy’ theory) actually predicts that there would be things we understand that you don’t. This means that the fact that you don’t know why we took you to the dentist’s is actually predicted by Alexa’s theory!

Of course, I would actually just tell Mikey to be quiet and go to bed. When he got older I would explain oral hygiene to him. But, I hope, the point remains clear: a theory doesn’t get explanatory power by literally taking things we didn’t understand and then helping us come to some knowledge about them. It gets explanatory power by predicting3 the things we see in the world. Sometimes, we look at things and think that they don’t make much sense on atheism, like the fine-tuning of the universe, so we say they call out for explanation. On theism, they are more likely, so we say theism explains them better. But it’s not like we actually went from not knowing something about fine-tuning to knowing that thing because of the help of the theory - unless that thing we learned to be true is theism itself. Instead, we started with facts we already knew, and looked for a theory that would predict them. Theories don’t teach us new facts about the things which they are predicting. Moreover, skeptical theism actually helps theism to predict what we see - it predicts that there would be evils we don’t understand, and that is what we observe.

Subscribe now

Denying Gratuity (TSTS’s first argument)

Moving on to the critiques that TSTS makes, the first is pretty simple. TSTS4 complains that skeptical theism would lead one to deny the existence of gratuitous evil or suffering altogether:

[…] skeptical theism appears to have some vicious side effects on one’s ability to interpret reality in a morally intelligible way. For one, it requires one to completely deny the existence of gratuitous suffering; one would have to believe that the fawn who burns in the forest fire died for some unknown, borderline indiscernible reason.

Perhaps it’s just me, but biting this bullet seems a serious blow to any coherent worldview […]

So far as I can tell, TSTS’s complaint here, if it is taken to be independent of the next point TSTS makes, is that the existence of gratuitous evil is just obvious and that denying it is silly and disrespectful to those that have suffered.

While TSTS is right that the denial of gratuity is a consequence of skeptical theism, I don’t think it is a problematic one. In fact, it is just a consequence of plain old theism. Any theist is already committed to the view that all evils are justified or outweighed5. Skeptical theism doesn’t create a new commitment here. As such, the claim that denying gratuity is silly isn’t an independent argument from the problem of evil. As for the claim that it is disrespectful, see the section on moral apathy in this article - the basic argument there is that an evil can be non-gratuitous even if it does not make the world a better place. As such, we can sympathise with people who have suffered and not diminish their pain even if we believe in skeptical theism.

No Falsifying Conditions (TSTS’s second argument)

TSTS argues as follows:

Under this, theism is unable to predict anything about the world and how it [could] be expected to operate; immense evils are expected, but so is no evil, effectively rendering the whole thing extremely unfalsifiable. No amount of data, in this case, could make a dent in the convictions of the skeptical theist. Such a sentiment reminds [me] of the debate Philip Goff had with Perry Hendricks, in which the la[tt]er unironically argued that, had we lived in a hell world, this situation would not constitute any evidence against God’s existence.

Now, I haven’t watched the entirety of the debate that TSTS mentions, so I am not going to pin my stripes to everything Hendricks says there. However, I have discussed this objection before in the comments under this post and do agree with the general idea of what Hendricks says. I will respond here as there:

While a hell world, with no future or past of goods connected to the hellish suffering, clearly constitutes evidence against God to us as external observers, that is because we are just that - external observers who know that the evils of this world are unjustified, ex hypothesi. However, if we were in that world and did not have the information that its evils were unjustified, because we didn’t know whether there were necessarily connected goods in the future, then we wouldn’t have such evidence against the existence of God. The ethical calculation for external viewers is simple and easy to perform - there is no good, and only evil, so this world’s evil is gratuitous. The calculation is not so easy for people within that world with limited knowledge of the future. If they were given some kind of crazy computer that could assess everything in their world, then they would have evidence against the existence of an omni-God.

As such, there are states of affairs which constitute evidence against the existence of God even under skeptical theism. It is just that we don’t see those states of affairs occurring (which is not a surprise for skeptical theists, because those states of affairs would only occur if God didn’t exist!).

Other Objections by TSTS

TSTS goes on to make the moral aporia/paralysis objection, which I addressed here.

TSTS also says:

Taking gratuitous suffering to be an illusion may also render our certainty about other, less apparent facts—such as the truth of a given religious faith or even scientific findings—to be unjustified. In activating this nuclear option, the skeptical theist gets burned, too.

I am not sure exactly what TSTS’s method is here to get from skeptical theism to skepticism about religion and about science but I have some ideas.

For the first worry, I assume that TSTS is referring to the worry that skeptical theism leads to skepticism about divine testimony because belief in divine testimony relies on the kind of inferences that skeptical theism blocks. In particular, one might think that belief in divine testimony relies on a noseeum inference like ‘I see no reason why God would lie, so I can conclude that He wouldn’t.’ Skeptical theism blocks such an inference. As mentioned in the first footnote, I am currently reading on this topic, so I will defer addressing the critique to that future post, and link to it here when I have posted it.

The second worry might be as follows: if God gave us faculties that are really bad at judging moral matters, then He also may have given us faculties that are bad at dealing with science and similar things. This is related to a worry given by Emerson Green in the podcast which TSTS links to6, so I think it is fair to interpret TSTS as referring to this. My response is as follows. The skeptical theist need not argue that God gave us poor moral faculties. In fact, many skeptical theists think we have good moral faculties for our own moral reasons. It is God’s moral reasons which we are impaired in cognizing7. This is because our own moral reasons are important for us to understand. God’s are not as important, because we don’t act on God’s reasons, we act on our own. We can therefore expect that scientific reasoning, which is so helpful, is something God would also give us the ability to understand.

Lastly, TSTS says the following:

Of course, I don’t demand that God dispense to us his every motivation and bit of knowledge, but I certainly don’t think it would kill him to throw us a bone either. Skeptical theism, it would appear, contends God can’t or won’t even do that much.

This objection appears only to come off if the skeptical theist happens to believe that there are no theodicies which work at all and they rely solely on skeptical theism to diffuse the problem of evil.

In fact, it need not come off even then: Murray points out in Nature Red in Tooth and Claw (2008) that some evils might be of the kind that telling us the reasons they occurred would undermine those very reasons: Murray gives the example of a parent who fools their child, hiding their tools in order to teach them a lesson about tidying up after themselves. If the child knew that the tools were hidden in order to teach them a lesson about tidying up, they would no longer believe that the tools were lost because of their carelessness, and so wouldn’t learn their lesson. Murray suggests that perhaps there are cases where God has reasons for allowing evils but revealing them would undermine the reasons, as in this example.

Finally, a committed skeptical theist could bite the bullet and argue that God’s reasons are so complex that He wouldn’t be able to explain them to us without massively intervening and changing us. Of course, this bullet-biting relies on the view that such an intervention is outside of God’s duty, and one might reject that idea. Nevertheless, this is a position some theists have held to and it is not absolutely unthinkable.

Share

Conclusion

None of the arguments given against skeptical theism come off, except perhaps the argument that it leads to skepticism about divine testimony, which I will address in a future post. The argument that skeptical theism is a theoretical failure misunderstands how theoretical reasoning works. The argument that it denies gratuity misses that that is just what theism says. The argument that it is unfalsifiable does not succeed because there are actually falsifying conditions. The brief closing arguments are also unsuccessful.

Skeptical theism is not an admission of defeat. It is actually successful!

Also, I’ve been thinking:

Caption: Do I need to caption this?

Alternative title: Everything Evil IV: Three Moral Problems for Theodicists

NB: This article is too long for most e-mail clients, so to read the full article, please click the title to open it in substack!

I’ve been sitting on this one a while. It fills the role of answering questions about moral scepticism, which I promised to address in my earlier article on the general scepticism challenge to sceptical theism, here, which was part of a series on the problem of evil, hence the alternative title.

However, Chris recently posted an article arguing that omission theodicies (and theodicies more generally) cause one of the moral problems I was planning to discuss. Thus, I have decided to reskin this post as a response to Chris. I had originally planned to read around a lot more before writing this, but one must strike while the iron is hot! I will update in the future if further reading changes my mind in any important ways.

The context

The problem of evil, as described here, argues that the existence of God is less likely than one might otherwise think, because of the existence of gratuitous evils.

Theodicies argue that evils which appear to be gratuitous are not gratuitous because they are either justified (in the case of the Free Will Defence, also detailed here, divine risk-taking theodicies, participation theodicies, and, according to Matthew’s comment here, omission theodicies) or otherwise outweighed by goods (in the case of soul-building theodicies, among others)1.

Sceptical theism essentially holds that there could be theodicies for evils which occur and appear to be gratuitous, but that it is reasonable to think we might not be able to figure these theodicies out.

I will, for brevity, term anybody who believes in theodicies, sceptical theism, or a combination thereof, a theodicist.

Theodicists thus hold that the gratuity premise in the problem of evil (the premise that there is gratuitous evil) is not warranted, or that we should have a lower credence in it than we might initially think.

In the remaining sections, I will address moral problems that arise from theodicies and sceptical theism. The first is moral aporia, which Chris argued for in his article, and is argued for in the literature by Piper. I will consider several responses, which I will categorise into three classes: failing responses, middling responses, and succeeding responses. In the failing class we have a contextualist response and one providentialist response. In the second we have a different providentialist response and a response which appeals to divine commands somewhat dogmatically. In the succeeding class we have the aforementioned distinction between justified evils and outweighed evils, and the relativist response, which I think is the strongest. Here, I will be engaging directly with Chris. I think Chris’ points fail against the relativist response. However, the relativist response leads directly to another problem: moral scepticism - Sharon Street argues moral scepticism is a natural result of this relativism. I will argue, as Pegan does, that it does not in fact follow. The third problem is moral apathy, an issue I anticipated and responded to on my website a few years ago2. I argue here, as there, that apathy only follows from a view of evil as exclusively outweighed and never merely justified.

Subscribe now

Moral Problem #1: Chris’ problem a.k.a. Moral Aporia

(Aporia is a fancy word for paralysis).

The worry that Chris and Piper3 propose is that for any evil we observe in the world, if we think there is a theodicy explaining it (or that there could be a theodicy we don’t know about, as in sceptical theism), then we should be paralysed or shouldn’t intervene to stop it, because the theodicy gives a reason to allow it.

Strictly speaking, Chris doesn’t say outright in his article that we should be paralysed or not intervene, he only says that there is a reason to allow the evil. However, I think the claim that we shouldn’t intervene is heavily implied by this line in Chris’ 3rd section:

“And so, according to this view, there is no problem with permitting Marvin from being assaulted.”

I think Chris is here implying that a human would resultantly be justified in permitting the evil. If Chris doesn’t mean this, then all he is really saying is that God has a reason to allow the evil, and that is just what the theodicist says, in which case I don’t see how this is a problem for the theodicist. Thus, I will take Chris to be arguing for the view Piper proposes, which is that the theodicist must be paralysed or not intervene.

Here’s a toy example I’ll use throughout to hopefully make things clearer:

Drowning boy: You are walking along the pier alone and notice a boy drowning in the sea. You can jump into the sea and save him.

In this example, Piper proposes that the theodicist should reason as follows: if I don’t intervene then God will have created a world in which the boy drowned. Per my theodicist views, this means that there is a reason to allow the boy to drown, because if there were no reason, God would not allow it. Therefore, at best, I don’t know if I should use my free will to prevent this putative evil (I am in a state of aporia), and at worst, I should not intervene.

This is obviously a problem for theism. Firstly, it appears blindingly obvious that you should intervene and save the drowning boy. Secondly, most theistic traditions include injunctions to prevent evils such as this from occurring. Thus, if the theodicist really is bound to this conclusion, this seems to be a strong reason to reject the theodicist arguments.

Let’s consider some responses.

Failing Response 1: Contextualism

4Here, I argued that a contextualist about knowledge has the resources to respond to an atheist who argues that sceptical theism leads to general scepticism. Perhaps we can follow that contextualist line here.

The contextualist would hold that when considering the problem of evil, much is at stake (you could doom yourself to hell or devote much of your life to somebody who doesn’t exist if you make a mistake), so the standards for knowledge are higher. Therefore, contextualism says, we should consider possibilities that are further afield than we might consider in everyday life. This includes theodicies and sceptical theism. As such, when considering the problem of evil, we should think the gratuity premise unwarranted5 because there are live possibilities which entail its negation.

In everyday life, however, the contextualist continues, the stakes are lower, so we don’t have to consider these possibilities and we are therefore justified in believing that there isn’t a reason to allow the evils we see before us. Thus, we should intervene and stop them.

This response is implausible insofar as you think contextualism is false. Furthermore, even if you think contextualism is true, you should probably reject this argument. In the drowning boy example, we need only consider the potential consequences to see this: if you decide not to save him when you should have, then an innocent person dies; if you decide to save him when you shouldn’t have (i.e. when God had a reason to let him die) then that boy could go on to do horrendous things or the boy could lose out on some brilliant goods that somehow resulted from his drowning (hence God wanting him to die initially). Either way, there are significant consequences, and so the stakes are high. Therefore, the contextualist should think the standards for knowledge are higher and we can’t rule out worrying possibilities, however marginal they might seem, like the boy going on to be an evil dictator. As such, the contextualist idea doesn’t work. We still find ourselves in moral aporia.

Failing Response 2: Naive Providentialism

6The thought here is that God might organise the world such that the right result occurs regardless of what one does. In the drowning boy example, this means that if God wants the boy to drown then He will providentially order things so that the boy drowns regardless of your intervention - perhaps despite your intervention, more waves arise, which sweep the boy away from you. If He does not want the boy to drown, He will order things so that you can save the boy. Therefore, we should intervene and do what we can to prevent evil, because if God wants that evil to happen, our efforts will have no effect. Since God is all good, He has probably organised the world in this way, the providentialist could argue, and so we need not experience aporia.

Yet it’s not obvious that this argument is one that most theists will want to make. Many theists are committed to the existence of free will and of a predictable law-governed universe. If humans have free will, then that means that God has to prepare for multiple scenarios, and if He prepares for them as I described above (e.g., the waves sweep the child away only if somebody tries to save the child), then that means the laws of physics are changing on the basis of my free decisions. Thus, the universe is not governed by predictable laws. If, alternatively, the theist denies that we have free will, so that God only has to prepare for one set of scenarios, they are already incurring costs, since theism is normally committed to the existence of free will. Free will can be a great help to theists in the free will theodicy for evil, so this a significant potential cost of providentialism.

There is another route concerning providence that the theist might wish to take, though. Let’s suppose they maintain that free will exists, but don’t think that God changes the laws of physics to account for our free actions. Instead, they claim that through meticulous planning, God has set things up so that our free actions, with the laws of physics operating normally, happen to set in motion causal chains that bring about the associated effects. With the example we have been using, this would mean that an attempt to save the child would, through the normal laws of physics, cause waves to occur that sweep the child away. This is a line of thought that the theist might pursue, but it seems prima facie incredibly unlikely that it is even possible for the universe to be ordered in this way - there are just so many complicated matters and possibilities of free actions that might be taken, that it seems wild to suppose God could set up the laws of physics just right to account for every single one. Indeed, even if it is possible that God set things up in this way, there may also be evidence against it. For example, if you have ever intentionally prevented an evil from occurring, only to later discover that things would have gone better if you had failed to prevent it, this might count as evidence against providence operating as suggested.

Middling Response 1: Molinism/Sophisticated Providentialism

7This response is somewhat different to the Naive Providentialism just discussed. Molinism is often known by a different name - Middle Knowledge. If God has middle knowledge, this means that He knows what you would do even in situations where you have a free choice over what to do. If this is true, then God’s plans do not have to be meticulous and all-encompassing as described above - God only needs to plan for one eventuality - the one which you will actually end up deciding to go through with.

This is only a middling response, rather than successful, for a few reasons. Firstly, Molinism is not without its own problems. Some argue that it makes God the author of evil, because He knew that people would freely choose to do bad things, and still actualised a world where they did so, even though He potentially could have created one where they didn’t (supposing you reject Transworld Depravity, as detailed here). You might think this is a problem for theism. Secondly, it is unclear whether middle knowledge really is possible if we take libertarian free will seriously - many people will think that if we truly have free will, there is no way for God to know in advance what we would do.

Middling Response 2: Dogmatism about Divine Commands

One line the theodicist can take is to be dogmatic about God’s commands. The thought here is that my knowledge that I should intervene to prevent evil, on the basis of God telling me to do so, in the Bible, the Qur’an, or through other revelation, is stronger than my knowledge that theodicies work, so I should just follow God’s commandments and intervene to stop the evil.

I labelled this as middling because of two flaws.

Firstly, the idea that my knowledge of God’s commands might be stronger than my knowledge of theodicies is potentially flawed - if the truth of theodicies just follows from God existing, and moral commands from God rely on the proposition that God exists and has spoken to me, then my knowledge of God’s commands must be less secure than my knowledge that He exists, and thus that theodicies are true. Thus, the central idea - that theodicies are to be believed less strongly than divine commands - seems implausible.

Secondly, if God’s commands to you are part His divine testimony, this argument needs it to not be the case that we should be sceptical of divine testimony as a result of theodicist views. It is a live question in the literature whether theodicism leads to scepticism about divine testimony. I have previously promised to write an article on this topic, and re-issue that promise here. When the article is out, I will come back and link it here. For now, it is enough to say that this response relies on a controversial view about divine testimony, and so it is not clear that it succeeds.

Succeeding Response 1: Distinguishing between justified evils and outweighed evils

This is the response Matthew goes for in the comment I linked to earlier. The idea is that just because an evil is allowed by God does not mean that it is outweighed by a greater good or by an evil that it prevents. It could be the case that it is merely justified, in the sense outlined in the first footnote of this article.

If an evil is merely justified, not outweighed, that means that it occurring does not make the world a better place. As a result, we are totally justified in intervening to prevent it from occurring.

I think this response succeeds, but its success is qualified: only some theodicies function such that they merely justify and do not outweigh evils. If Matthew is to be believed, the omission theodicy is one such theodicy. I am not in a position to judge that question. Nonetheless, I would be surprised if it were the case that every evil out there were merely justified and not outweighed. Put differently, I am saying that it is plausible that some evils are outweighed by greater goods. As a result, this is only a partial response to the moral aporia worry: we should intervene to prevent an evil if we think it is merely justified and not outweighed. But this does nothing to help us if we think the evil is outweighed by a greater good or prevented evil, and if we are unsure about whether an evil is justified or outweighed, then we should be unsure whether to intervene on the basis of this argument. Thus, I think we should go in for a stronger response, which follows.

Succeeding Response 2: The Relativity of Reasons

This response is due to Sharon Street8. Street points out that reasons for intervening to stop evil can be both evidence-relative and agent-relative.

Let’s imagine you have a two-year-old son, and he is about to trip over a toy he has left on the floor.

Evidence-relativity: You can’t see that there is a pillow which will cushion his fall. I can see it. In this case, you have a reason to intervene to prevent the putative evil – your son’s head hitting the floor. I, on the other hand, with the evidence regarding the pillow, do not have as much of a reason to intervene. As such, reasons can be evidence-relative.

Agent-relativity: You have told your son multiple times before to be careful about leaving that toy lying around, and you have decided that in order to teach him a lesson, you will let him fall and hurt his head. Being his parent, you have the authority to make such a decision about how his life will go. I do not. Thus, I have a reason to intervene, while you do not necessarily have one. The difference is because you are the kind of agent who has authority over your son’s life, and I am not. Thus, reasons can be agent-relative.

To return, then, to the drowning boy, God might have reasons to allow that boy to drown, be they evidence-relative, or agent-relative, while we, as humans without all the evidence and authority, do not have those reasons. God is the creator of the universe, so He has certain authorities we do not have, and God is omniscient, so He has some evidence that we do not. Thus, we should intervene, even though God might not have the same obligation.

Here is an excerpt from my aforementioned article which makes the point clearer:

The moral aporia argument crucially relies on the idea that the bare possibility that there are some reasons to let an evil happen constitutes a reason not to intervene, even if we don’t know what those reasons are. This is clearly not true. An example given by Bergmann & Rea in ‘In defence of sceptical theism: a reply to Almeida and Oppy’ (2007) is very good at making this point:

Suppose you witness a child being kidnapped and know that they are likely to die if the kidnapping succeeds, but you can intervene. It is possible that the child is actually a monster who will destroy the world if you save them. Nonetheless, it is clearly the case that you should save the child, and even if it turned out that they were a monster, we would not say you actually had a reason to let the child get kidnapped. It was just that there was a reason, but it was one you didn’t have.

Thus, acknowledging the relativity of reasons gives the sceptical theist a very strong way out of the moral aporia objection.

Now to engage directly with Chris. Chris anticipates this relativist response and says the following:

First, the specific claim I am making is that if there exists some relation of justification J for God g to permit some evil e (Jge), then more generally, there exists some justification to permit some evil (∃xJxe). And this is a perfectly normal existential introduction. Putting the critique like this illuminates its inescapability–in submitting a theodicy which justifies God permitting evil, the theist cannot escape the generalisation that there exists justification for evil.

I would respond: Chris, you are right. But there being a reason to allow the evil to occur does not imply that we humans have a reason to allow the evil to occur.

Chris has another response:

Second, introducing God’s knowledge as a symmetry breaker between what is justified for God and what is justified for people is a classic red herring. Recall that God is justified in permitting Marvin’s assault on the fact that it builds his character–note how the justification is grounded on this fact and this fact alone. This entails that knowing this fact generates justification for permitting Marvin’s assault. And this fact is surely something that people can know without God’s divine knowledge (indeed, it certainly must be known by those submitting a theodicy, otherwise what would they be submitting?).

Sure, if we knew that an evil were linked to a theodicy, then we would have the same evidence as God and so would have the same reason not to intervene which God has. However, I don’t think we do always know that an evil is linked to a theodicy. Perhaps some theodicies are committed to this, but certainly the general theodicist, for whom sceptical theism is a possibility, need not know about the theodicy which is involved. In fact, the sceptical theist claim is that we do not know! So, we need not have the same justification that God has.

Also, having the same evidence as God doesn’t address agent-relativity problems.

Subscribe now

Moral Problem #2: Moral Scepticism

Even though she proposed the relativist response, Street argues that it leads us into moral scepticism. Let’s see why she thinks this: to begin with, Street says that some of her other work establishes that there is no good realist secular moral epistemology. This has the result that if we are to insist on moral realism (as many theists do), then we are forced to accept that God has some role in helping us know about and understand moral truths. Herein lies the rub, Street argues: if reasons are agent-relative, and God’s reasons are not known to us by standard reasoning, as the sceptical theist argues (and other theodicists might think), then there is a problem. Our knowledge of our own moral reasons – the moral facts – relies on God having a reason to help us. If we don’t know that He does have such a reason, we cannot think that our moral knowledge is reliable. Thus, Street claims, the theodicist’s appeal to agent-relativity undermines their knowledge of morality.

However, this challenge from Street has a fatal flaw: the claim that in order to think our moral knowledge is reliable we must first believe that God has reasons to help us know about it. I do not think this is true, even if one thinks that our moral epistemology necessarily involves God, as I hinted at the end of this article.

Instead of needing to believe that God has reasons to help us, we are justified in holding our moral beliefs through the normal method by which we come to believe them. We don’t also need to make claims about that method’s reliability. If, of course, we had good reasons to believe the method to be unreliable, that would be a problem, but the bare possibility that it is unreliable does not itself give us a reason to believe that it is. As a result, believing that we don’t know God has reason to help us know about morality does not mean that we don’t know about morality9.

To see this clearly, think about how we normally reason morally - we look at the case, and think about whether it would be morally justified to intervene or not, and then settle on one view. We do not precede the consideration with ‘Aha! I can reason about this, because God helped me to do so, because He had a reason to help me do so! So my moral deliberation is reliable and I can do this!’ No, we just do the moral deliberation, and it is reliable (because God helped us - if you buy into that view). We don’t need auxiliary hypotheses about God’s reasons just to deliberate on moral matters.

You might think this is all fine, but we have a defeater once somebody points out to us that we couldn’t possibly know that God had a reason to help us. I don’t think this line works. Firstly, I don’t believe that somebody pointing this out constitutes a defeater - our moral beliefs are prima facie justified regardless of a sceptical theist’s inability to discern God’s reasons, and secondly, I think we can know that God has a reason to help us, independently of inferences that a sceptical theist would reject:

If we think that God has told us that He helps us know about morality, perhaps by saying that He has written the law on our hearts, as He says in the Bible, then we can know that God has a reason to help us know about morality, one which doesn’t rely on any sneaky noseeum inferences or other suspect methods that a sceptical theist has ruled out. We should notice, though, that this route relies on us not being sceptical about divine testimony, as mentioned earlier. The same thoughts I gave there will apply here.

Moral Problem #3: Moral Apathy

The final challenge to the theodicist is moral apathy. The challenge runs something like this: if God had reasons to allow an evil which occurred in the past, then we shouldn’t lament the occurrence of past evils - they were all for the greater good. We should be morally apathetic. This seems like a cost to the theodicist, since there are events which seem very clearly lamentable, and this argument potentially bars theodicists from lamenting evil.

I think this argument fails because it misunderstands what it means for God to have a reason to allow an evil. It supposes that God allowing evils is always a straightforward case of goods outweighing evils. However, this need not be the case. Let’s see why (you will already have an idea if you have read the first footnote of this article, or read Succeeding Response 1 to the first moral problem).

Some theodicies argue that certain evils which God allows are outweighed by goods. As an analogy, we might think the suffering we go through when we go to the dentist is outweighed by the good of our teeth being healthier. This seems like a case where we shouldn’t lament the evil (suffering) because it is outweighed by a good (health).

However, not all cases of God having a reason to allow evils are cases of outweighing -- they are not all analogous to the dentistry case. Some are cases of justification, not outweighing. For example, the Free Will Defence is a well-known theodicy which can account for the existence of moral evil. If one believes that God grants free will to demons, then it might also account for natural evils such as avalanches and volcanic eruptions, which might be thought to have been caused by demons. The free will theodicy doesn’t declare that the evils which occur when agents exercise their free will are strictly outweighed by goods, but that they are justified10. The possibility of us using our free will to perform morally good actions justified us having free will, and thus justified God allowing evils to occur. The Free Will Defence is not the only theodicy to function like this - any theodicy that involves God taking risks in His government of the universe will function similarly. As mentioned above the participation theodicy also involves justification, and Matthew argues that the omission theodicy does too. These theodicies allow us to lament evils, since the evils are not being outweighed by goods, but justified. It is right to lament that an agent making use of their free will did the wrong thing, even if we can recognise that God had a reason to permit the evil. Thus, we need not be morally apathetic if we believe the Free Will Defence and similar theodicies. As a sceptical theist, we are ambivalent about whether the unknown theodicies involve outweighing or justification, so we can also reject moral apathy.

It is also important to note another route that a theist might take in responding to the moral apathy challenge - biting the bullet. They might agree that they shouldn’t lament any evil that occurs in the world, for exactly the reasons that the challenger gives. They may think that the only reason this seems like a challenge to their view is because the challenger does not properly appreciate the fact that all the evils they see are leading to greater goods. I myself do not find this line of thought convincing, but it is certainly a route that is open to the theodicist. This is, compared to moral aporia and moral scepticism, a relatively small bullet to bite.

Conclusion

To conclude, a theodicist need not fall into moral aporia, moral scepticism, or moral apathy. The argument for moral aporia relies on the false premise that the bare possibility of there being a reason to permit evils undermines our reasons to stop the evils. The argument for moral scepticism relies on a false premise about how we come to know facts of morality. The argument for moral apathy relies on a misunderstanding of what it is for God to allow an evil - it supposes that God allowing evils is always a straightforward case of goods outweighing evils. However, this need not be the case.

Have you ever seen a spider egg sac open, releasing hundreds of tiny spiders to roam free? It sends shivers down your spine, right? Or have you seen somebody leg-pressing far more than they should and extending their leg so much that their leg buckles and their knee starts bending the wrong way under the force of the weight? The idea makes me feel sick. Lastly, have you ever been swimming in the ocean before suddenly realising that there is a huge abyss below you that could contain anything from stinging jellyfish to a giant kraken? There’s a sinking feeling in your gut at the thought.

These were some of the examples I gave at a philosophy debate club to argue that aesthetic value is objective. You see, many people had spent a decent chunk of the night arguing that beauty must be subjective because people disagree about what is beautiful. So, I thought I could offer some instances of negative aesthetic value which everybody would agree on, and thereby give some evidence to the view that it was objective.

This was an awful argument.

Not only are the examples not unanimously agreed upon - somebody who handles spiders for a living probably does not get freaked out by baby spiders - but the unanimity of a position wouldn’t prove objectivity. If everybody who disliked the colour blue died, leaving only blue-enjoyers alive, that wouldn’t mean that blue was objectively a pleasant colour.

Now, of course, unanimity can be evidence for objectivity over subjectivity. If you asked 1000 people a question in a survey and they all gave the same answer, this can be evidence that you asked a question which had an objective answer like ‘Does water boil at 100°C?’ rather than that you asked a question with a subjective answer like ‘Do you like tea?’

The problem comes in when people think that agreement automatically, without any other considerations involved, shows objectivity, or, more frequently, that disagreement automatically shows subjectivity. In debates I have often heard that some cultures think cannibalism is permissible, so morality must be subjective, or that my friend might find The Ugly Duchess attractive, so beauty must be subjective. This is clearly silly. The shape of the Earth is not subjective, but there is disagreement over that. The existence of God is not subjective, but there is widespread disagreement over that. If you are going to use disagreement to argue for subjectivity, you’ll need a lot more argumentation, and you’ll have your work cut out for you1.

#6 Disregard thought experiments because ‘that would never happen’

Like with the previous example, there are times when this is perfectly reasonable. For example, in debates over moral responsibility this move can be seen:

Incompatibilist: Determinism is incompatible with moral responsibility, because of the Principle of Alternate Possibilities - to be morally responsible for a decision, it needs to be the case that there were genuinely possible alternatives to the decision you did in fact make.

Compatibilist: I disagree. Consider this example of a Frankfurt case: You are about to assassinate the innocent man, Bob. Unbeknownst to you, there is a chip inside your brain which I control. The chip has the effect that, if you were ever to back out of assassinating Bob, I would be able to press a button that would force you to kill him anyway. You then decide to kill Bob, and I never use the chip. There was no alternate possibility where you didn’t kill Bob, but you’re clearly still morally responsible for Bob’s death regardless. In a Frankfurt case there is moral responsibility despite the lack of alternate possibilities.

Incompatibilist: Aha! But Frankfurt cases are impossible. Either the brain states which the chip detects determinately require you to kill Bob, in which case there are no alternate possibilities, whether or not there is a chip in my brain, and so no moral responsibility; or the brain states don’t determinately require me to kill Bob, in which case the chip can’t accurately predict whether I am about to kill Bob, and the chip can’t serve its purpose.

This is perfect. The Frankfurt case is, according to the incompatibilist’s argument, truly metaphysically impossible. This means that the compatibilist can’t use it as an example of a case that contradicts the PAP.

What won’t do is something very similar, but not quite the same, which is when you don’t show that a thought experiment is impossible, but that it is incredibly unlikely. Here are two examples:

When giving a presentation in a tutorial on the topic of a priori knowledge, I considered a thought experiment by Hawthorne2. Assuming that for knowledge to be a priori, it must be indefeasible by experience, Hawthorne has us consider something like the following:

Imagine you believe that ‘All bachelors are unmarried’. You then are imprisoned in solitary confinement for 10 years, in which time the use of the word ‘bachelor’ changes to such an extent that it is no longer correct to say that ‘All bachelors are unmarried’. You then leave solitary confinement, and learn that this semantic change has occurred. Through your experience, the belief that ‘All bachelors are unmarried’ has been defeated. Thus, your knowledge which appeared to be a priori was not in fact a priori, according to the definition we are considering.

Now, whatever problems you might think there are with this argument, it is clearly wrong to say the argument fails for the reason I gave: ‘this thought experiment is incredibly unlikely to occur, since usually you are part of a linguistic community when it makes semantic changes like that.’

Of course, Hawthorne’s whole point is not that this sort of thing happens often, but merely that it is possible! The mere possibility that the supposedly a priori belief could be defeated by experience is evidence that it is not a priori, because we said for something to be a priori, it must be indefeasible by experience - that is, it must be impossible to defeat it with experience.

Usually thought experiments don’t need to be even remotely likely in order to give us the force required to make an argument come off. They just need to be possible (either logically or metaphysically, depending on your views about modality).

If you have looked into epistemology, you will no doubt be familiar with the brain-in-vat thought experiment. Here, you are asked to imagine that you are a brain in a vat, being fed nutrients and electro-chemically manipulated by a mad scientist, such that you, the brain, believe that all sorts of normal things are happening around you and to you. Crucially, the brain in the vat has all the same appearances before it as you do in the real world, though they are not veridical.

Granting that this thought experiment is possible would suggest that we can’t be 100% sure of any of our beliefs about the external world, because it remains compatible with our appearances that we are in fact brains in vats. Now is not the time to refute scepticism (for a post that details some ways to deal with scepticism, see here), but we should note that it is entirely unsatisfactory to respond like this:

Sceptic: If you could be a brain in a vat, then —

Me: Ah! But that’s a BIG IF!

Sceptic: What?

Me: Why would a neurosurgeon want to put me in a vat? That’s just so unlikely! It would serve no purpose; it would be so time consuming, blah blah blah …. You get the idea.

The sceptical argument doesn’t at all rely on brain-in-vat being likely. It just needs to be possible in order to (supposedly) undermine your knowledge of the external world.

Subscribe now

#5 Demand definitions for all words

The late great Robert Nozick has the following account of knowledge in his Philosophical Explanations3. For a proposition to be known by a person, it must be the case that:

1. It is believed by that person.

2. It is true.

3. If it were not true, they would not believe it.

4. If it were true, they would believe it.

We might be unsatisfied by Nozick’s 3 and 4, thinking the conditions for their truthfulness are ambiguous or vague. Thankfully, Nozick gave us a brief analysis of conditionals to help us understand 3 and 4, saying that 3 basically means ‘in possible worlds close to the actual world where the proposition is not true, the person does not believe the proposition’. Similarly, 4 means ‘in possible worlds close to the actual world where the proposition is true, the person does believe the proposition’.

However, unsatisfied with even this analysis, I gladly announced in one of my undergraduate philosophy classes that Nozick’s analysis must fail because the word ‘close’ in his analysis of the conditionals is hopelessly ambiguous: what makes one world closer to another - similar propositions? similar atoms? similar states of affairs? This ambiguity, I thought, could not be disambiguated, and so Nozick’s account could not help us understand what knowledge was.

This goes wrong in two big ways. Firstly, there is a pretty large literature on what it takes for one world to be close to another, of which I was simply not aware. This is true of many philosophical terms - in a debate, you probably won’t be able to pick out one single word that doesn’t have a clear definition,thereby making the whole view hopelessly ambiguous - philosophers have and can get more precise about words like this. Secondly, you often don’t need to get precise for a view to be intelligible and defensible. It’s true of many words which we understand that we don’t have a good definition for them. This doesn’t stop us using them in intelligible ways. You (presumably) wouldn’t say that a definition of hydrogen as an atom with one proton is a lacking definition if the person proposing it can’t quite say what a proton is. Or that a child saying ‘bachelors are unmarried men’ who can’t also give a very precise definition of marriage is therefore making some kind of grave error. So why do this to philosophers? It is true that one of the great things about philosophy is getting precise about difficult terms, but that doesn’t mean the burden is always on a philosopher to give a definition of a term in order to use that term. In my argument against Nozick, I had confused not having a reductive analysis of a term with not understanding it at all.

#4 Accuse everyone of begging the question

Begging the question is typically understood as assuming the conclusion you’re trying to prove. For example, if I argued as follows, I’m clearly begging the question:

1. If my intuitions tell me something is true, it’s true.

2. My intuitions told me that if my intuitions tell me something is true, it’s true.

3. (Conclusion, From 1 + 2) If my intuitions tell me something is true, it’s true.

Begging the question in this very narrow sense creates terrible arguments, because you would never be able to convince anybody of 3 using this argument. If they already doubted 3, then they would doubt 1 (since 3 = 1), so the argument wouldn’t sway them.

However, outside the very narrow set of cases where a premise is identical to a conclusion, it becomes a lot harder to tell when arguments are question-begging. As a result, lots of beginners in philosophy end up accusing people of begging the question when they are doing no such thing.

One example that I think about often was when I was confronted with the zombie argument for property dualism and it struck me that the argument begged the question. The argument (or at least part of it) runs:

1. Phenomenal zombies are possible.

2. If phenomenal zombies are possible, then property dualism is true.

3. Thus, property dualism is true.

Spurred on by a friend, I thought: why would anybody believe that phenomenal zombies are possible if they didn’t already think that property dualism was true? Of course, property dualism must be true if phenomenal zombies are possible (that is to say, of course 2 is true), but that means that only people who are property dualists already should accept 1! So the argument begs the question.

The problem here is, again, twofold. Firstly, it seems that I am complaining about the argument being logically valid: the second premise is true, and so the truth of the first premise entails (and is entailed by) the conclusion. Framed as I framed it above, this sounds like a problem, but, as others have pointed out, this is exactly the definition of a deductive argument! You want your premises to entail your conclusion. So, something must be going wrong here if this is what begging the question amounts to.

Secondly, it is just wrong that believing property dualism is the only reason to believe the first premise. Chalmers and other proponents of the argument give reasons to believe premise 1, namely that phenomenal zombies are ideally conceivable and ideally conceivable things are possible. So, my opinion that the argument begged the question largely relied on ignorance of part of the argument. That part I wasn’t thinking about does require more philosophical argumentation, but it makes for a case that I really don’t think begs the question.

In my example, I confused the fact that the conclusion follows from the premises with the notion that the premises are justified only by assuming the conclusion. Those are two different things. The latter is question begging, but not true of the zombie argument. The former is true of the zombie argument but is not question begging - it is exactly what we want from an argument.

If you want to read more about what it takes to actually beg the question, there are some wonderful posts out there: see, for a start, here and here.

#3 Talk only about language and never about reality

This comes up a lot when we’re arguing about whether various things are subjective or objective, but it can also occur elsewhere. The trick that you can accidentally perform on yourself (and on others) is to point out that interaction with reality is mediated by concepts and language, and then act as though the reality thus does not exist or is not important.

One interaction my friends and I had with Timothy Williamson when we invited him to our student society struck upon this problem. Williamson claimed that we could make genuine advancements in knowledge, giving the example of us learning that whales aren’t fish because they give live birth rather than spawning. My friends and I worried that actually what happened was that we re-classified whales as mammals, making a change in our language - how we spoke about whales - but that there was no underlying change in our knowledge. This (rightly) frustrated Williamson!

The mistake here is that there are two phenomena - a change in knowledge, which resulted in the other - a change in language. It is tempting to treat the change in language as though it is the only real change and to ignore the change in knowledge which preceded it.

This is closely related to a similar problem with thinking about language - thinking that the concept is the most important fact and somehow trumps ideas about mind-independent reality. For example, we might be tempted to think that prior to certain scientific advancements water was not H2O but, only upon the scientific advancement, it became H2O. Of course, the truth is that water is and always has been H2O, regardless of what we know about it. Here is a funny clip of Vaush seeming to make this mistake and say that in Mexico water is not H2O:

If we think of language as our map of reality, it is important to note that our map can change without reality changing, but, often, we change our map because we learned something new about reality.

Share

#2 Be a general sceptic when it suits you

When somebody presents a philosophical argument to you for a conclusion that you don’t like, it is very easy to bunker down and attack any and every premise with ‘and how do you know that?!’

I have done this myself many times. One example was when arguing against the problem of evil, I have insisted that the atheist doesn’t know the premise about the gratuity of evil and so can’t rely on it. I would become extremely sceptical when this premise was proposed, and bring up all sorts of sceptical scenarios like the aforementioned brain-in-vat thought experiment. This felt like I was making a good response, but I really wasn’t.

Firstly, most people aren’t committed sceptics, so they don’t think that the putative possibility of brain-in-vat and similar scenarios do not in fact undermine our knowledge. Thus, using these scenarios might stun somebody uninitiated in the field of epistemology, but if you know what you’re talking about, this is dishonest - very few philosophers really think that we must be sceptics because of these scenarios, regardless of whether we know how to respond to them. It is likely that you don’t actually believe in scepticism yourself, so you shouldn’t argue like you do just to bring down somebody’s premise when you don’t like it.

Secondly, if you are a genuine general sceptic, like I used to be, it is still odd to put forward these arguments - you needn’t know that a premise is true to use it in a good argument. It might just be that we reasonably believe a premise or have a high credence in it, and so we should update in favour of the conclusion of the argument. I was perfectly happy to walk around as a general sceptic saying that I reasonably believed that my feet wouldn’t phase through the ground when I took a step, even if I didn’t know that, but all this charitableness disappeared when I saw somebody make an argument I didn’t like: suddenly it became all or nothing - you KNOW or you DON’T! This is unfair.

#1 Obsess over fallacies

Lastly, and most annoyingly, to suck at philosophy you can become an obsessive about fallacies. The different ways people can make mistakes in reasoning are no doubt interesting and fun to analyse, but as actual problems to diagnose in people’s philosophical thought, they are much less useful than advertised.

In his work on utilitarianism, J. S. Mill argues that the idea of justice is a useful fiction that humans have invented in order to maximise utility. I accused him of committing a genetic fallacy - arguing that a proposition is false because of the reasons some people have for believing in it. You might think that this is an accurate description of what Mill was doing, and so commend my accusation of his using a fallacy, but I think this is a very uncharitable construal of Mill. Mill might be more rightly described as proposing an error theory - he argues that justice is not real in earlier chapters, and here explains why people nonetheless believe in it - because it is a useful fiction. Read in this way, Mill is clearly not performing a fallacy.

In light of this, and as proposed by Maarten Boudry here, a dilemma comes up: either we can make a fallacy’s definition very broad so that we can capture real arguments people make into it, but then it becomes useless, because those people are not clearly making a mistake in reasoning, or we can make the definition very tight in order to make it clearly a mistake, but then the fallacy becomes vanishingly rare in real practice.

Instead of focussing on all the ways reasoning can go wrong, it is much more useful to learn the ways it goes right - the laws of logic, rules of statistical inference, and perhaps some other inferential methods. Note that knowing when things are going right immediately allows you to notice when things are going wrong, too! Instead of plastering a name on the mistake, you can just point out that the argument doesn’t follow the laws of logic, or that the non-logical inference made isn’t valid. This takes much less work from you (learning all the different fallacies is exhausting!) and is far less frustrating for your interlocutor: if they have spent a long time arguing for a conclusion, then identifying one or two premises as somewhat similar to a fallacy that somebody has named in a book, and thereby thinking you have proven the argument wrong, without even showing that a premise is unjustified or an inference invalid, then you are the one arguing badly here - not them! The mere fact of an argument or premise being named as fallacious doesn’t make that argument unsound or that premise false.

For something to be morally good or bad, it has to affect a person. This seems like an obviously true ethical principle, right? Not to modern moral philosophers! The past few decades of moral philosophy have seen rigorous debate over the ‘Person Affecting Restriction’1 thanks to the rise of population ethics, a field concerned with actions that affect which people and how many people exist in the future. Population ethics is complicated stuff, but in this article I want to examine a more basic ethical principle inspired by debates in population ethics. I’m calling this more basic principle the ‘Experience Affecting Restriction’. The principle seems prima facie incredibly strong, but there are several ways in which it has been challenged, varying from considerations about everyday intentions to grand cosmological theories.

The Experience Affecting Restriction (hereafter EAR): Nothing is morally good or bad if it does not affect the experience of an existing thing.

This restriction is not as strict as the Person Affecting Restriction (hereafter PAR), because it spans over all things that experience, not just persons. This means that if you don’t consider animals to be persons, you can still consider actions like vivisection or factory farming to be morally bad. This specification is also wide enough that if you thought there were things that could experience but weren’t really alive like, say a conscious artificial intelligence, you may also consider things that affect them to be morally important. Note this specification doesn’t even require that the experiencing thing has pleasure or pain as part of its experiences. Maybe it’s possible to have experiences that are morally important but which have nothing to do with pleasure or pain. If so, the EAR accounts even for them.

So, thanks to its incredibly wide-ranging account of moral importance, the EAR actually restricts very little – there’s not much that it says is morally unimportant. Thus, it seems very difficult to contest. Nonetheless, a surprising number of thinkers throughout history have raised ideas that serve as challenges to the EAR.

Non-Identity Problems

The PAR became a contested notion in recent moral philosophy thanks to considerations in population ethics, primarily those raised by Derek Parfit. Parfit raised examples like:

The 14‐Year‐Old Girl. This girl chooses to have a child. Because she is so young, she gives her child a bad start in life. Though this will have bad effects throughout this child’s life, his life will, predictably, be worth living. If this girl had waited for several years, she would have had a different child, to whom she would have given a better start in life2.

Many of us have the intuition that in The 14-Year-Old Girl, the right thing for the young girl to do would have been to delay pregnancy, so that her future child can have a better quality of life. However, it is odd to take the well-being of future persons into account when evaluating actions. Doing so in a completely unconstrained way can lead to strange problems. For example, if immoral events like war or the rule of corrupt political regimes led to certain future people being born that wouldn’t have otherwise existed, we may be forced to say these events had morally positive consequences3.

It’s important here to note two possible interpretations of the PAR:

PAR(1): Nothing is morally good or bad if it does not affect an existing person.

PAR(2): Nothing is morally good or bad if it does not affect a person, existing or future.

Now, there are many considerations on either side of this interpretive divide, but the cases we’ve considered thus far give us some motivations either way. If we accept PAR(1), ruling out future persons from the morally relevant set of persons, then we will avoid the problem cases where we’re forced to call war good. However, this comes at the cost of sacrificing our intuitions about The 14-Year-Old Girl: accepting PAR(1) forces us to say that the girl isn’t doing anything wrong. Thus, you might prefer to accept PAR(2), embracing the moral relevance of future persons. To support this move, you might argue that the cases of war and corrupt regimes do genuinely create moral goods, but these goods are small compared to the wrongs involved. If you reject consequentialism, you may also very easily deny that wars and corrupt regimes are responsible for the lives of the relevant future persons.

Now, PAR(1) is closer to the EAR than PAR(2). If you accept PAR(2), then you think that future persons are morally important, and future persons fall outside of the range of the EAR. Thus, if we are to pick PAR(2) over PAR(1), to save our intuitions about The 14-Year-Old Girl, then we should also reject the EAR. Thus, despite the EAR seeming incredibly intuitive and obvious, population ethics gives us our first reason to doubt it.

If successful, however, this argument may be seen as a call to modify the EAR rather than to abandon it. It is not too much of a change to the principle to expand it to include future things as well as existing things. The following considerations in the rest of this post conflict with the spirit of the EAR much more clearly and would be much harder for it to accommodate.

Subscribe now

Intentions

Another reason to doubt the EAR concerns events much more widespread than those in population ethics. Intentions, which occur almost all the time, are things to which we often assign moral importance, yet may be seen as disconnected from anyone’s experience. Suppose this thought experiment:

The Meteorite Murderer. A man sitting at home contemplates murdering his wife. He decides he is going to get up and murder her, and neurones fire in his brain, sending signals to his legs to stand up so he can go about the murder. However, before those signals even reach his legs -- before his intention has been actualised in any way -- a meteor strikes his house and completely destroys him.

It seems as though the man’s intentions are immoral, even though they never had a chance to affect any person or indeed any experiencing thing. This intuition might be particularly convincing for those who subscribe to certain virtue theories which ground the rightness or wrongness of an action in the intentions behind it4.

There are several ways we might save the EAR in response to this consideration, though. Firstly, we might argue that the intuition that the Meteorite Murderer’s intention is immoral is a misfiring intuition – the only reason we have this intuition is because in practically all cases such intentions actually do lead to actions that affect experience. Thus, our intuitions are trained to consider such intentions as bad, but this thought experiment plays on that tendency in a deceptive way. Secondly, we can point out that just by virtue of a person having certain intentions, those intentions affect the experience of a person – the person who has the intentions! Thus, this isn’t really a violation of the EAR. Initially it seems very odd to hold that murderous intentions are morally important because of the way they affect the person who has those intentions, but such a view may also be more plausible to virtue theorists, who see the intention as making the person who has it into a less virtuous person.

Thus, intentions might militate as counterexamples to the EAR, though there are clear counterarguments. Yet if the counterarguments fail, and intentions really are violations of the EAR, then this shows that the seemingly obvious principle is violated over and over again in our daily lives.

Desires

Desire-satisfaction theories of well-being posit that a person’s life goes well when certain desires of theirs are satisfied. Success Theory is one such view considered by Derek Parfit5, one which is distinctive in that it considers all desires about one’s own life to be important, not just those which are introspectively discernible. For example, if you desire not to be lied to, Success Theory will count the satisfaction of this desire to be good for your well-being even if you have no way of telling that it is in fact being satisfied. This is in contrast to other desire theories which might reject the not-being-lied-to desire as irrelevant, since from your internal perspective there is no difference between the case where it’s satisfied and the one where it isn’t.

Success Theory poses a threat to the EAR. Since well-being is usually considered morally important, the truth of Success Theory would suggest that there are morally important states of affairs which are important regardless of their being experienced. For example, suppose that you have a desire to be a good parent. Unfortunately for you, there’s a nuclear explosion, which kills your son, who is in another city, and then kills you several minutes later. For those several minutes, a state of affairs obtained where you were no parent at all, so your desire was not being satisfied. If Success Theory is true, this state of affairs is probably morally important. Arguably, though, this state of affairs was never experienced by anyone, because your experience would have been just the same whether or not your son had died before you did. Thus, this seems like a counterexample to the EAR.

However, this argument relies on a controversial understanding of what it means to experience something. You couldn’t tell the difference between the situation where you died first and the one where your son died first, but that doesn’t mean that you never experienced the state where your son died first. Quite clearly you did experience that state, because you lived through it. You wouldn’t say that nobody ever experienced being asleep just because they couldn’t tell the difference between being asleep and being awake. As such, we shouldn’t say in this case that you never experienced the state of affairs where your son died first. Thus, this argument fails to establish a counterexample to the EAR.

Beauty

In Section 50 of his Principia Ethica, Moore argues against Sidgwick’s version of the PAR. Sidgwick had declared anything that didn’t affect people to be outside the realm of ethics. To support this view, Sidgwick argued that it would be bizarre to pursue beauty if it couldn’t be appreciated by any people.

In response, Moore asks us to consider two worlds entirely devoid of any creatures to observe them: one filled with beauty, nature in perfect harmony with itself; the other filled with horror, a completely disgusting mess. He affirms that if we had to pick between actualising one or the other, even though we nor anybody else would ever observe either world, we should pick the beautiful world. Perhaps you share Moore’s intuition, and so think that, even if it is only a matter that’s usually of little importance, beauty actually is morally valuable, regardless of its being experienced.

However, even if you share Moore’s intuition, this thought experiment faces a problem: in asking us to imagine the two different worlds, isn’t Moore asking us to imagine ourselves experiencing the two different situations? It seems impossible to actually posit a thought experiment that doesn’t involve any experience, since by imagining the situation, we imagine ourselves experiencing it. Thus, maybe Moore’s attempt to get at an intuition that disfavours the EAR falls flat due to the nature of thought experiments.

Perhaps Moore could reply by asking us instead not to imagine the two worlds, but to imagine ourselves in a voting booth making the decision, knowing that nobody would ever get to see the worlds. However, in such a case, the intuition which Moore wants us to feel seems a lot weaker. In addition, it’s not clear that we could have such knowledge about the worlds without anyone ever coming into any contact with them. Thus, the thought experiment might not even be possible6. If Moore’s appeal to intuition succeeds, this is a very clear violation of the EAR, but it’s not at all clear that his appeal is successful.

Share

Theism and the EAR

It is interesting to note how often the EAR is challenged when it comes to the philosophy of religion. However, as I will explain, there may in fact be something special about theistic views, something that leads them to accept ideas which look like violations of the EAR but actually aren’t…

In Nature Red in Tooth and Claw, Michael Murray explores several arguments that protect theism from the problem of evil, particularly applied to animal suffering. In exploring these arguments, he raises (not one, but two!) ideas that seem to violate the EAR.

Firstly, Murray explores Neo-Cartesianism, the view that animals do not suffer in the same way as humans, and that this difference is a morally relevant difference, such that animal suffering isn’t really an evil to be counted in the problem of evil. An opponent of Neo-Cartesianism might argue that this can’t be true, because if it were, then it would be permissible for humans to wantonly abuse animals. Since this isn’t permissible, Neo-Cartesianism isn’t true.

However, as Murray points out, a theist might be able to maintain the impermissibility of such wanton abuse while maintaining the Neo-Cartesian position. They might do this by positing moral laws which are counter-examples to the EAR: maybe it is immoral to upset the proper functioning of nature, independently of the experience that animals may or may not have. For example, a theist may consider it immoral to destroy sand dunes that naturally protect land against erosion. Moral principles like this might seem egregiously ad hoc attempts to save the Neo-Cartesian position, until we consider the support that Abrahamic religious tradition has given to the view for many centuries. God considered the world to be ‘very good’ (Gen. 1:28) before He created human beings, and, as Murray points out, the Bible describes God as ‘entering into covenantal relationships with the natural world and as engaged in the project of redeeming it (Gen. 9:8-17; Rom. 8:18-22)’. Of course, for such a view to be truly successful, an account of what ‘nature’ really means would need to be given, something which is notoriously difficult7. It also seems that the view might only be convincing to those who believe the Bible, but nonetheless there is clearly promise here for a properly staked out position that rejects the EAR.

Secondly, Murray explores an argument surrounding animal suffering that relies on a different putative rejection of the EAR: the view that the world moving from chaos to order (CTO) is intrinsically good. I won’t explore the success of this argument here, but instead the underlying principle. Similarly to the considerations surrounding the proper functioning of nature, Murray details that there is substantial support in religious traditions for the view that CTO is good in and of itself. Figures including Gregory of Nyssa, Augustine, Descartes, and Henry Ward Beecher have all posited such a view. Unfortunately, it’s not easy to offer arguments in favour of the intrinsic goodness of certain states except by asking you to consider intuitions in thought experiments. Beecher considered the idea of somebody creating one huge watch-making machine to be clearly more magisterial than the idea of that person going about making many watches bit by bit. He likened the one huge watch-making machine to the CTO process of evolution, which, according to His view, God set in motion. However, these thought experiments seem to have a lot to do with majesty, and not so much to do with morality. Perhaps none of these views ever intended to speak on moral goodness, but some wider kind of goodness. I’m not sure what this wider goodness might be, though8. Nonetheless, regardless of the kind of goodness under debate, the use of thought experiments may still fail for the same reasons we considered against Moore’s thought experiment.

Even if we consider the two views posited by Murray to be correct, there’s another reason to doubt that we’ve truly found violations of the EAR, and that is the very theistic framework in which the views are found. If God exists, He presumably observes absolutely everything that occurs, or at least knows about it, and so anything that occurs affects His experience. Thus, neither the moral goodness of nature’s proper functioning nor of CTO can truly be considered violations of the EAR if we believe that God exists – absolutely nothing can fail to affect the experience of an all-knowing God, and so on theism, nothing can exist that is both morally important and doesn’t affect anyone’s experience. This is the aspect of theism I mentioned earlier, which may lead theists to support views that look like violations of the EAR, but actually aren’t.

There is one more potential violation of the EAR which falls within a theistic framework, and arguably doesn’t suffer the problem just mentioned. Axiarchism is a theistic view, most famously posited by John Leslie, but which doesn’t involve the same kind of God as in Abrahamic religions. In his 1989 book Universes, Leslie looks at various pieces of evidence for fine-tuning in the universe. If the evidence for fine-tuning is not being misunderstood, it shows us that if the laws of physics were randomly determined, the chances of them taking life-permitting values are ridiculously low. These low probabilities cry out for explanation. Typically, philosophers explain the low probabilities by appeal to the God of a traditional religion, who chose these values in order to bring human beings into existence. Others explain the probabilities by positing a multiverse, with each universe containing different laws of physics – ours just so happens to be the one with life-permitting laws. Leslie, however, explains the low probabilities by positing axiarchism – the view that ‘some set of ethical needs is creatively powerful’9.

The idea here is that the universe came into existence in the way it did not through some complex multiverse process, nor through the work of a traditional theistic God, but because it was morally necessary that it do so. This moral necessity is presumably derived from the fact that if the universe didn’t exist in this way, persons would not, and the existence of persons is a moral good, or allows for moral goods. However, there must then be some other explanation for the vast swathes of universe which are, as far as we can tell, unnecessary for the existence of persons. This is where the EAR is potentially violated. If Leslie’s view is correct, the most plausible explanation might be that it’s morally good for certain parts of the universe to exist, even if nothing ever experiences them. Of course, Leslie’s view is easily the least plausible of all the EAR violations considered thus far, but those interested in axiarchism should read W. J. Mander’s chapter on the subject in his Idealist Ethics. Mander details Leslie’s theistic ideas, which the reader might think leads it into the same territory mentioned above, where EAR violations are impossible because God experiences everything. Thus, even Leslie’s view might not truly lead us to accept any violations of the EAR.

Conclusion

So, despite the strong intuitive initial appeal of the EAR, there are several arguments that might convince a few of us to reject it. All of these arguments have big problems, so none of them should be considered binding and there probably isn’t even a strong cumulative case against the EAR. Nonetheless, it is at least interesting to consider whether the domain of ethics might rightly be extended beyond experiencing things, and what motivates some thinkers to do just that.

Religion is the opiate of the masses. - Karl Marx

When I was 13, I became an atheist. 5 years later, I became a theist. In both instances, Marx’s idea of religion played a critical role. In both instances, my arguments sucked.

Why I became an atheist

I became a teenager in 2016, the year of the Brexit referendum in the UK and Trump’s first election. Because of these events (and because a lot of celebrities randomly died that year), I remember lots of people saying it was the worst year ever. I knew practically nothing about politics, so I didn’t really get what the fuss was about.

In an attempt to figure out what all this political stuff meant, I started looking into political ideologies. If you’re a young person reading this: not necessarily a good place to start - most politicians in the 21st century don’t subscribe to a neatly written out ideology. Nonetheless, this is where I began, and, of course, what other ideology to consider first but communism? (Do your facepalms here).

I picked up a very short book on the life and writings of Karl Marx from the school library, and stayed up late at night reading it. As soon as I read the Marx quote at the beginning of this piece, the scales fell from my eyes and I could see clearly: belief in God was motivated primarily by feelings - the wish to avoid death, the desire for justice, and for purpose in life. Ultimately, I thought, there was nothing else to it - nobody had ever presented me with an argument for God’s existence. As a child I had read books answering questions about God and laying out that theism was consistent. But showing the consistency of the view did not show that it was warranted. So, it seemed that belief in God was entirely unwarranted. I concluded that theism was an irrational position, which was probably my first genuinely philosophical opinion. That’s where I left my consideration of theism for a few years as I continued looking into politics, which I thought was far more important.

Subscribe now

Why I became a theist

I didnt’ realise it at first, but in my fumblings for a political ideology, I was looking for something like a purpose in life, or a value around which my life could be organised. In communism, that was equality. Later on in my thinking, as a socialist, that value was equality along with the good of the common man.

Yet none of this really rang true at a gut level. For much of my life as a teenager I was very sad1. Maybe the cause of the sadness was the fact that I was a teenager with hormones all over the place, or perhaps it was social media, or something else. I’m not sure. However, I can tell you exactly what I thought the problem was at the time: I felt, once I put political ideas aside, that there was no point to my life. I was looking for something that would motivate me to get up in the morning, which would give me a reason to put up with everything that bothered me in life. What was the point of doing my schoolwork? It might get me a job and make me happy, but what was the point of my happiness? Why aim at it? What was the point of anything? I know now that this problem, if you can call it a problem, is very much a first world one, but for me at the time, it was a problem nonetheless.

When I was 17 or 18, I became more interested in faith, at first because I had a crush on a girl who was a Roman Catholic, but then more in earnest. I noticed that my mother had a copy of C. S. Lewis’ Mere Christianity, which I read with great interest. None of Lewis’ explicit arguments convinced me of theism, but one which he seemed to repeatedly hint at was the one that sealed the deal for me, and which related to my struggle over purpose. Here it is:

1. If God doesn’t exist, there is no purpose in life.

2. If there is no purpose in life, nothing we do can be ultimately wrong.

3. (From 1 + 2) If God doesn’t exist, there is nothing ultimately wrong with believing He does.

As a result, I thought, one could believe in God merely because one wanted to have some purpose in their life. You might be fooling yourself - there may be no actual purpose in reality - but that wouldn’t matter anyway.

I became a theist in Summer of 2021. Later that year I inherited a book from my grandmother, C. S. Lewis vs the New Atheists by Peter S. Williams. There I found several more and much stronger arguments for the existence of God, along with a section where Williams makes explicit the argument I had sensed in Lewis’ work: in addressing arguments for atheism, he writes ‘How can anyone be convinced to adopt a worldview that denies that anyone should ever be convinced of anything?’2

To return to Marx: if the masses were fooled into believing that God existed without good evidence just because it made them feel as though their lives were more meaningful, I was now one of the masses. To think myself distinct from the masses and better than them, able to discern purpose and meaning without God, was folly. As says:

If my belief in God was epistemically unjustified, so be it. Why should epistemic norms matter in a world with no God?

Share

Commentary on the arguments

The Marxist line on theistic warrant

I now know that there are plenty of arguments for theism, and that not all of them (almost none of them) say ‘balls to rationality if atheism is true!’ like the one I have just described does. Among others, the fine tuning argument stands out to me as convincing. documents fifteen here. If one thinks any of the arguments work, one need not agree with Marx’s line.

Moreover, even if none of the arguments work, I don’t think one has to throw their lot in with Marx and believe that theism is always unwarranted. You may have an idea of what I am about to say from the fact that I keep referring to warrant - that is the term used by Alvin Plantinga for the property that turns mere true belief into knowledge. Plantinga’s famous book, Warranted Christian Belief, argues that if God exists, theistic belief of the usual kind is warranted. That is because God gives us a ‘sensus divinitatis’ or ‘divine sense’, which is put to work when we experience great beauty and sense God’s handiwork, when we sin greatly and feel God’s anger, when we repent and feel His mercy, etc. To show that this doesn’t happen, Plantinga argues, one would need to first show that the divine sense didn’t exist, and this is, he thinks, pretty much the same task as showing that God doesn’t exist3. Even if the theist doesn’t have any of the good arguments for God to hand, their belief can still be warranted, just as long as God actually does exist. For Marx to be right that theistic belief is unwarranted, it would have to be the case that God doesn’t exist, and Marx is in no business of proving that.

So, because I believe God does exist, I think the Marxist line on warrant is mistaken.

The Lewis-style argument for theistic belief

The theistic argument is very unclear in the form I have written it above. Alas, that is how I thought of it when I was 18. I will try to clarify what I meant at the time, and along the way discuss why it goes wrong.

First, let’s try to demystify Premise 1. ‘Purpose’ seemed like an intelligible idea when I was young, but I am now less convinced that it is. I think the best approximation of what I meant by ‘purpose’, was something involving norms (moral, epistemic, aesthetic). Norms could give me something to aim at which genuinely mattered.

However, I wanted the sense in which it mattered to be objective, rather than grounded in feelings or opinions. This was part of the motivation for Premise 1 - I thought subjective normativity didn’t require God, but it was unsatisfying. Objective normativity, on the other hand, would give me purpose, but required God.

Now, this seems like a huge error on my part. Why think only objective norms can give somebody purpose in life? Sure, purpose provided by living up to subjective norms won’t matter to the universe, but I see no reason to think it must. The opponent of the argument can say ‘The only person to whom my actions need to seem purposeful is me! I don’t care what the universe thinks!’ And to me, that sounds like a perfectly respectable position. So, the claim that purpose requires objective normativity, and therefore God, is a mistake, or at least not dialectically productive.

Let’s move forward. I’ve said that I thought objective normativity required God, but why did I think this? It had something to do with grounding. I was convinced that normative facts had to be non-natural in some way. I think I was swayed by the ‘just-too-different’ intuition, documented by David Enoch4, which says that moral facts (or all normative facts) are just too different from natural facts for them to be reducible to natural facts. So, if normative facts existed they had to be non-natural. And if they were non-natural, the thinking went, they had to be grounded in something else non-natural. I was not yet familiar with abstract objects such as numbers, propositions, or properties, and so I didn’t like the idea of having any free-floating abstracta in my ontology. Therefore, grounding normative facts in some non-natural object seemed like the only way to preserve them. The only believable (to 18-year-old me) non-natural entity which could ground normative facts was God. And so, the only way to ground normativity was by appeal to God’s nature. Thus, we have Premise 1. As we can see, I only got to it via a bad argument about objectivity, a largely uninterrogated intuition against normative naturalism, and an ignorance of abstracta.

Let’s quickly clarify Premise 2: By something being ‘ultimately wrong’, I meant that it was, all-things-considered, not what you should do by light of the norms mentioned above. I think this premise is true, given the understanding of ‘purpose’ I have just outlined. However, one true premise does not a sound argument make.

Finally, a little note on the conclusion: it doesn’t strictly allow that belief in God is permissible! It could be the case that God exists, yet that there is no good reason for believing in Him. It would then be impermissible to do so. This is compatible with the conclusion, so the conclusion doesn’t actually give an iron-clad reason to believe in God like I thought it did. Thankfully, Plantinga’s work on the sensus divinitatis gives very strong reason to think that if God does exist, there is reason to believe in Him, as do arguments for God’s existence.

Conclusion

Although neither of the arguments I have presented are convincing, hopefully they can show people on both sides of the apologetic divide how not to argue about God. From my time in the field I have seen people making the same mistakes as me in both directions, but especially in arguing for theism. If you’re an apologist reading this, please don’t argue like I did!

Although the theistic argument doesn’t work, there is an argument in the vicinity that I think shows more promise. If you buy into a meta-ethics like the one outlined by David Enoch in this interview, I think a moral knowledge argument for the existence of God is going to be rather convincing. If there is interest, I’ll say more about this in a future post.

1. It appears that we are designed to kill animals to eat their meat.

2. If it appears that we are designed to Φ, it is not always immoral to Φ.

3. (Conclusion, From 1 + 2) It is not always immoral to kill animals to eat their meat.

Share

Defence of Premise 1

Many aspects of human biology suggest that we are designed to eat meat, alongside other foods1. One example is human teeth. Your canines appear designed to tear through flesh. Another example is your digestive tract, which is shorter than a herbivore’s and longer than a carnivore’s, reflecting an omnivorous diet. Your gut produces all the necessary enzymes to digest meats. Some nutrients are incredibly hard to find naturally except in meats. And, frankly, meat is delicious. If we were not meant to eat meat, we could expect it to be disgusting to us (or at least not delicious) like many things we shouldn’t eat are.

The Roadkill Objection

Objection: What you have said doesn’t show that Premise 1 is true! Premise 1 said that we are designed to kill animals to eat their meat, while the evidence so far merely shows that we are designed to eat meat. What it doesn’t show is that we are designed to kill the animals. Perhaps our biology is designed so that we can find roadkill and animals that have died of old age to eat.

Response: Eating roadkill or animals that have died of natural causes doesn’t appear to be part of the design. Firstly, such meat is often unhygienic, having rotted in ways that would prevent a human from consuming it without risking disease. Animals that die of natural causes obviously do not receive the same treatment that slaughtered animals do, and they therefore are less appropriate for eating. Secondly, humans that ate only meat of this kind would only very rarely find meat that they could eat, and so the digestive system’s readiness for decent quantities of meat does not accord with the theory that we were designed only to eat roadkill or relevantly similar meats. Thirdly, humans are clearly designed to hunt as well as to eat: we have eyes set in the front of our heads, rather than the sides, allowing for binocular vision, a feature that many predators share and prey do not. We are able to run and walk extremely long distances, which is terrific for hunting. We are also able to track animals effectively and use tools to hunt them. These last two points are not very strong evidence of design for hunting, as they are arguably just results of our ability to reason (rather than being capabilities specifically designed for hunting), but nonetheless they may be convincing points to some.

Subscribe now

The Genesis 9 Objection

Objection: Christians shouldn’t think we are designed to kill animals for meat. It was only in Genesis 9:3, as part of the Noahide Covenant, that humans were given meats to eat:

Therefore, before this covenant, humans ate only plants, so they can’t have been designed to kill animals for meat.

Response: Humans are designed to feel pain and grief, but, according to many traditional views, before the Fall there was no physical harm to elicit pain or death to elicit grief. Therefore, it is evident that on Christianity, God’s design reaches beyond the initial stages of the world, and anticipates later developments. Thus, the fact that the very first humans didn’t do certain things doesn’t mean we are not designed to do those things. It appears that God anticipated the Noahide Covenant by designing humans to eat meat even if they didn’t do so until the Noahide Covenant.

Defence of Premise 2

I find this premise intuitive, and my hope is that it will be accepted by people regardless of their philosophical stripes. However, I recognise that it is a lot more believable to theists than it is to atheists. This is because a theist will see human design as the result of a moral agent who seeks the good, and so is unlikely to create humans designed to do something immoral.

The Skeptical Theism Objection

Objection: Adam, you are a skeptical theist, as you have shown in other posts on Substack. This means you believe that you don’t know what reasons God might have for allowing certain evils. You therefore should be willing to accept the possibility that God might design something in a way that is in fact immoral. Thus, it’s possible that God could have designed humans to kill animals for their meat, despite it being immoral to do so.

Response: You’re right, I am a skeptical theist. Luckily for many of my readers, they may not be, and so this objection doesn’t work against them. Nonetheless, I have a response available to skeptical theists: firstly, as clarified in this post, my skeptical theism is not the view that we can’t know what reasons God might have for allowing certain evils, but is instead the view that we can’t come to know about those reasons via noseeum inferences: we can’t reasonably argue ‘I don’t see the reasons, so they ain’t there!’

Objection: So, what?

Response: Well, we can still know via other methods of inference that God wouldn’t allow certain evils. In this case, it seems we can know that God didn’t allow this evil - the evil of immoral design - because God explicitly allows the consumption of meat (in Genesis 9:3 above). God saying that He will allow you to do something is as good as a confirmation from God that doing that thing is not immoral2.

The Aliens Objection

Objection: Some wasps are designed to burrow into human eyes and lay their eggs there. Premise 2 would establish that such behaviour is morally permissible.

Response: Non-human animals are not (usually) moral agents, so actually their behaviour is always morally permissible.

Objection: Okay, but we can imagine aliens that are moral agents and behave in a similar way, like the aliens in Aliens. Surely it is not moral for them to kill humans in order to reproduce? It seems, then, that Premise 2 is false.

© 2017 Twentieth Century Fox Film Corporation. All Rights Reserved.

Response: Premise 2 uses the term ‘we’. Unless my readership happens to include aliens, the use of the term ‘we’ is limited to human beings, so the premise does not apply to those aliens. Therefore, the example does not show Premise 2 to be false.

Objection: That’s just gerrymandering! The point is that a similar premise should apply to aliens who are moral agents if it is a genuine moral fact. So, is it moral for these aliens to reproduce in the way they do?

Response: This is a problem for many ethical views, not one unique to Premise 2. For example, virtue ethicists who ground ethics in human flourishing are subject to a similar attack, because it seems that alien flourishing involves killing humans. There are a couple responses that virtue ethicists give, one of which might work here: we can say that it is not just gerrymandering to apply the principle only to humans, because human morality is not inter-species - it applies only to humans and to try to apply it to aliens is a category mistake. This will not be convincing to everybody, especially those with a realist bent, because many of us have the intuition that if something is immoral, it is immoral for any moral agent, regardless of species.

However, we don’t need to take this virtue ethical approach. Instead, we can point out that Premise 2 initially got a lot of support from theism, and theism also points us to a response to the Aliens Objection. The idea here is that God would not create a universe containing such aliens, because God would not design moral agents to do something immoral. This is one of the reasons we believed Premise 2 in the first place, so it is not an ad hoc response to the Aliens Objection.

Thus, it appears to me that the Aliens Objection should only persuade you if you are an atheist moral realist and not a virtue ethicist.

The Subjectivity Objection

Objection: Whether or not something appears to be the case is subjective! Thus, somebody could employ Premise 2 to justify pretty much anything to themselves. For example, they might say that it appears to be the case that we are designed to murder people and thus, by Premise 2, that it is not always immoral to murder. It clearly is always wrong to murder, so Premise 2 must be false.

Response: The sense in which I use the word ‘appears’ is not one which is subjective. In the sense I am using, one must present evidence which suggests a certain conclusion about design. There being evidence of this kind is what it means for it to ‘appear’ to be the case that we are designed for something.

Objection: Then why use the word ‘appears’? Why not just say ‘If we are designed to Φ, it is not always immoral to Φ.’ and give your evidence for that premise?

Response: The sentence you suggest is only acceptable to somebody who believes human beings genuinely were designed. Somebody who doubts intelligent design would not be able to accept such a premise. By using the word ‘appears’, the argument encourages assent from anybody who thinks that evolution gives rise to features that appear like design, but are not actually designed.

David Enoch is a Professor of Philosophy and Law at the Hebrew University of Jerusalem. In October, he’ll be starting as the Professor of the Philosophy of Law at the University of Oxford1. He has written extensively on meta-ethics, and his book Taking Morality Seriously: A Defense of Robust Realism, argues for non-naturalist moral and normative realism. In this interview, I discuss the arguments David makes in his book and in one of his papers, ‘Why Idealize?’

Adam: What does it mean to take morality seriously?

David: Maybe I should start with a disclaimer. I don’t claim that fellow metaethicists who are not realists or who are not robust realists of my kind, don’t take morality seriously or are bad persons or anything along those lines. The thought expressed in the title of the book, however, is that if we fully flesh out the commitments that are involved in taking morality seriously, then some real metaethical implications follow.

Taking morality seriously means two things. Firstly, we should accord moral considerations weight in our deliberation. Perhaps sometimes we should do the right thing, even when it fails to promote our own interest or when we dislike doing it because of other reasons.

Secondly, there’s a touch of an interpersonal thing going on here. To take morality seriously involves, among other things, the willingness to apply moral judgments to other people as well as to oneself. In popular culture, sometimes people like to say ‘Don’t judge’ and sometimes that’s a good recommendation, but not always. It is sometimes important to judge, and there are metaethical implications to that as well.

Adam: So, what are the details of your view, robust realism, which you say follows from us taking morality seriously?

David: On robust realism, there are moral and more generally normative truths and facts. They are, in whatever sense we can make of such locutions, part of the fabric of the universe; they are perfectly objective. Their existence or truth does not depend on us in any constitutive way. When we are successful in our normative inquiries we discover them, rather than create, or construct, or legislate them. And furthermore, their nature, these moral facts or properties are not what are often called natural facts or properties. Natural facts and properties are very roughly the facts and properties that are of the kinds that are studied by the natural sciences. Moral facts are very different from those, but they are not in any way less real for that.

Adam: You’ve done a lot of work in law and the philosophy of law. There seems to be a similarity between your meta-ethical views and certain ways of thinking about the law - there are objectively true facts about what people should do, and they can’t be reduced to any other facts. Do you see any analogies or disanalogies between law and your metaethics?

David: Well, in fact, I’ve written a paper in which I argue that the legal case is strongly disanalogous to the moral one. The main point, I think, is something like the following. Morality is essentially normative; it’s essentially reason giving. You ought not to be indifferent to morality. That’s for me about as close to bedrock as you can get. You may think of this as an aspect of taking morality seriously.

On the other hand, the question whether you should be indifferent to law seems to me to be entirely open and contingent. Maybe if your country is quite decent then maybe you shouldn’t be indifferent to law. You should think of the law as giving you reasons for actions. Real reasons, not legal reasons or fictional reasons or would- be reasons, but real reasons, things that genuinely count in favor or against actions. But if my country is in a horrible state and it’s terrible in its policies and in its constitution and legislation and so on, then it’s quite possible that the law just doesn’t give me any reasons for actions at all. Or anyway, it doesn’t in virtue of being law. Maybe I should take it into account because I don’t want to be put in jail or something. But that’s a different kind of normativity. So I think that whereas morality’s normativity is a starting point, law’s normativity is anything but. And I argue that the fact that morality is intrinsically normative is a part of what makes metaethics interesting.

One other difference, a closely related one, is that we can change the law in ways that we cannot change morality. The law is either entirely or at least to a large extent an artifact. It is something that we collectively can shape. I don’t think there’s anything analogous to that phenomenon when it comes to morality.

Adam: Let’s go into some of the arguments you’ve made for the objectivity of morality, then. You’ve argued that the practice of standing one’s ground in moral disputes tells us something about objectivity in ethics.

​David: Yes. There are going to be conflicts between people about what we should do. For example, both you and I want the last cookie - I want the cookie, you want the cookie, there’s a conflict. We can’t both have all we want, right? In some contexts, it’s okay to just act on your desires, but there are contexts, like this one, in which we should really step back and recognise that there’s something deeply symmetrical about the situation: There’s no sense in which my desires are more important than yours, intrinsically as it were, nor is there a sense in which yours are more important than mine. A better, more morally relevant way to describe the situation is ‘David wants a cookie and Adam wants a cookie’, rather than ‘I want a cookie and someone else wants it’.

So, a symmetrical solution is called for. We can, for example, flip a coin, or I can say “Adam, you enjoy the cookie; next time I’ll have the cookie.” The thought here is that there’s a kind of equality. Roughly, it’s all about you and me and our preferences. My preference doesn’t count for more than yours.

However, there are other kinds of conflict, those that are grounded in moral disagreement. For example, we might have different opinions about dog-pain and its moral significance. Maybe you think that dog-pain counts morally as a reason against a particular course of action, and I don’t. Now suppose we have to decide about a joint action of ours regarding the pain.

Does this conflict seem just like the cookie conflict? Is it just about your preference for one kind of action, and my preference for another kind of action? If so, we should step back and go for a symmetrical solution like flipping a coin. Or is this a case where (if you can without significant further costs) you should stand your ground and say “No, no. What makes the difference here is not that it’s my preference or desire against dog-pain. It’s not about me at all. It’s about the dog, the pain, and the moral significance of that pain.” It seems clear to me that in these cases, at least to the extent that you can, you should not step back. You should stand your ground.

Now, I’ve drawn this distinction - conflicts that are based on moral disagreement behave differently from conflicts that are based on mere desire or interest or preference. In the former set you should stand your ground, but you should step back in the latter. Then I ask, “What explains this difference?”

If you go for some subjectivist view, according to which moral views just are expressions of preferences, interests, desires, or passions, then it’s going to be very hard for you to explain this difference. It seems that you should step back in all conflicts that are based on mere preferences.

I argue that the best explanation of the distinction is that when we’re engaged in a moral disagreement, it’s not about our preferences at all. There’s an objective matter of fact that plays a role. That’s the best explanation of why it is that we should stand our ground in some cases of moral conflicts based on moral disagreement. And, so, if this argument works, we have some reason to believe that there are objective matters of fact in morality.

Adam: There’s an ethical premise here in your argument for the view that ethical statements are true. It seems like a lot of people are likely to think this is question begging. You’re saying we should stand our ground in moral disagreements, and we shouldn’t stand our ground in cases of mere preference. This premise, combined with others, will get us to the conclusion that moral statements are true. But, if you assign a truth value to the first premise, which is a moral statement, in order to make this argument, you’re already committing yourself to the view that some moral statements are true. How would you respond to this kind of criticism?

David: First of all, I fully acknowledge that there is a moral premise there. I’m assuming on intuitive grounds that you should step back and view the situation symmetrically in the cookie case, but you should stand your ground in the case that’s based on a moral disagreement. So yes, there is a moral premise there. Does it beg the question?

As I grow older, I become more and more worried about begging the question in general. You think initially that it’s a fairly small and recognizable kind of fallacy that we understand well, but it turns out that it’s way more complicated than that. I mean, there’s this joke that philosophers like telling:

“Obviously your objection to my view begs the question because my view was not supposed to have any objections against it. So, clearly you are begging the question against me.”

It turns out to be really not at all simple to understand. But here’s something that I think is true about begging the question. Whether an argumentative move begs the question depends on the context of conversation. It depends on who the intended audience is. I want to say that at least vis-à-vis the vast majority of interlocutors, it doesn’t beg the question. And then, I have something to say about the remaining few.

Subscribe now

Adam: And those few would be error theorists, who just say there is no truth in the moral sphere?

David: I think that it would be a subset of error theorists, yes, not those who say there is no truth in the moral sphere, but those who deny the legitimacy of first-order moral discourse. But most meta-ethical theorists are not error theorists, and even many error theorists are perfectly happy to engage in first-order moral discourse.

They don’t differ with me about the legitimacy, as it were, of engaging in moral discourse and practice. Rather, they differ with me about how to best understand what it is that we do when we’re engaging in moral discourse and practice. So with those people in mind, it seems like I’m entitled to engage in first-order moral discourse. So I’m not begging any questions against them. Of course, they are free to reject my moral premise on first-order moral grounds. Maybe they’ll think that in the cookie case, I should not step back. I should just go as quickly as I can for the cookie. Or maybe they’re going to say that in the dog pain case, I shouldn’t stand my ground. I should just compromise, and to hell with the dog. But I don’t find these views very plausible. And for what it’s worth, the book has been out there for quite a while. These arguments have received at least some attention. I don’t know of anyone who seriously plays that game. The first order judgments that I’m relying on are fairly minimal.

So as long as I’m having a conversation with someone who is happy to acknowledge that it’s okay to do morality, not meta ethics, morality, right? Even if they’re not realists, even if they’re not my kind of realists, then I’m entitled to rely on first-order moral premises. So with those conversations in mind, I feel very confident to say that I’m not begging any questions.

Adam: Aren’t there some meta-ethical views, though, where they’ll say that we can engage in moral discourse, but moral discourse doesn’t have truth value? They might agree with you that impartiality does hold in some cases, and doesn’t hold in some other cases, but they wouldn’t assign a truth value to that kind of statement. They’ll just say it’s sort of an expressivist statement.

David: Yes. So of course, these expressivists are going to say “I can play the moral game as well as anyone else. In particular, just as well as the moral realist. But I have this fancy other story about what we’re doing when we are engaged in moral discourse. We’re not expressing beliefs, we’re expressing some very different kind of attitude.” They sometimes say a conative rather than a cognitive one. Whether they’re also going to say that moral judgments don’t have truth values, this is something that will vary among different expressivist groups.

We can imagine this conflict in a dialogical form. So I say, wait, do you agree? I’m going to be this annoying Socrates, right? “Surely, Adam, surely you agree with me that in the cookie case, we should step back and go for a symmetrical solution.” And then you have to be this Socratic figure where you just say “Of course, David”. And then I say something like “Well, but surely you agree with me that in the case of the dog pain, that, at least if I can, I should stand my ground and prevent you from causing unnecessary pain to the dog. I should not go for a symmetrical solution.” Notice that all I’ve done so far is first-order normative ethics.
And then I can ask, “Do you think that there is an important difference, a moral difference between how we should treat the disagreements and conflicts in the cookie case and the dog pain case?” And again, they’ll say, “Sure, there’s a difference.” And then I can ask, “And what explains this?” And I think they still need to explain it. So I’m not sure that going expressivist solves or somehow de-problematises the situation. That challenge arises for expressivists just as it does for anyone, as far as I can see.

Cookies and dogs: are these two things alike?

Now the interesting question is whether because they are expressivists or because they are a quasi-realist or whatever, whether this makes some solutions to the challenge available to them that are not available to other methods. That’s the more interesting path. I don’t see that it does. Let me just give you one example of a kind of a strategy of coping with a challenge that may be more open to at least some expressivists.

Some of them will be able to say something like the following: Sure, I agree with you that you should step back in the cookie case, but you should stand your ground in the moral case, but I don’t agree with you that I owe you an explanation of this fact, because either it is not fact, there are no moral facts, moral judgements don’t express facts. Or maybe it is a fact, but it’s a fact in a kind of attenuated sense. It’s not the kind of fact that gets into strong explanatory relations. And then the question goes back to how plausible it is to say such things and whether expressivism has other problems that they have to deal with and so on.

Adam: So, what about the error theorists we mentioned earlier, who think first-order normative discourse is illegitimate?

David: So some people are error theorists about morality, which roughly means that they think that morality is systematically erroneous. There’s a kind of error or mistake that runs through the whole of moral discourse. Most of us are error theorists about astrology. It’s not just that we think that a specific astrological judgment is false. We think that there’s some deep systematic error, and we think that this error just shoots through astrology. There isn’t a single astrological statement that I’m willing to assert because it’s just ridden with error in a systematic way.

If you are an error theorist of this kind about morality, you will not be willing to assert a single moral statement. In particular, you will not be willing to assert that it’s wrong to stand your ground in the cookie case, or that it’s wrong to step back and go symmetrical in the dog pain case. And against those people, I will be begging the question. But also, then I will fall back on how complicated begging the question is. And I would insist that there’s a sense in which I’m begging the question, but it’s not a terrible sense.

Share

Adam: So, this sense in which you’re begging the question, but it’s not so bad - would you say something like this is true? - The intuitions in favour of your first-order moral premises are stronger than the intuitions in favour of error theory, and therefore, even if you beg the question against error theory, the error theoretical views against which you’re begging the question are actually just weaker on intuitive grounds, so it’s not so bad. Is this how you would characterise the question-begging?

David: Yeah, I think that’s at least very close. Yeah. It does go a little bit deeper, but that is very, very close. The point is not just about how much intuitive support one version has or the other, because I think that the way in which we should and to an extent do evaluate competing philosophical theories is holistic. So you don’t just judge them for intuitive appeal. You judge them for intuitive appeal and explanatory power and how neat and beautiful they are, and maybe some other, maybe, you know, just give me the list of theoretical virtues and that’s how you do it in this holistic kind of way. But I do think that there are going to be cases in which begging the question is unavoidable.

And then you should just go at least provisionally with what seems to you to be true. I say provisionally because maybe further down the argumentative road, you’re going to come up against objections that are going to render the competing view more plausible. And then you should switch. But at least initially you should do that. I like to quote David Lewis’ work on logic on this. “No truth does have, and no truth could have, a true negation . . . That may seem dogmatic. And it is: I am affirming the very thesis that Routley and Priest have called into question, and – contrary to the rules of debate – I decline to defend it. Further, I concede that it is indefensible against their challenge. They have called so much into question, that I have no foothold on undisputed ground. So much the worse for the demand that philosophers always must be ready to defend their theses under the rules of debate.”2
When push comes to shove, when it’s that basic, there is no alternative to going with what seems plausible to you. In David Lewis’s defence, in my defence, that’s exactly true also of the other side. It’s not as if the other side has the strength of argument, wisdom, and reason on their side. And we mainly insist on what seems plausible to us. In whatever way in which David Lewis begs the question against Graham Priest, Graham Priest begs quite a few questions against David Lewis. In whatever way I beg the question against this kind of error theorist by relying on a moral premise, the error theorist is begging the question against me by insisting that I not rely on a moral premise. So at that point, we are stuck. And that’s where you just have to go with what seems to you to be true, in a provisional way, subject to future defeat and so on. So, I’m with Moore here, with the famous Moorean proof about the existence of an external world against the skeptic or against the idealist.

Adam: You’ve also made an argument for metanormative objectivity, as opposed to metaethical objectivity, from considerations about deliberation. Could you outline that argument?

David: Let me just start by briefly saying something about the relation between the moral and the normative. Moral considerations are the ones that come up in ethics about what’s right and wrong and maybe good and bad and so on. The category of the normative, I understand it as a wider category.

Realism about normativity is a logically weaker thesis than to be a realist about morality. To be a realist about normativity is to insist that there are at least some perfectly universal objective normative truths. But that doesn’t mean that those have to be in the smaller circle of morality. They can be somewhere else in the normative sphere.

This second argument, the indispensability argument, is an argument for the weaker thesis, for realism about the normative. And then when you combine it with the argument we just discussed, and maybe with other considerations, it can also support moral realism. But officially this argument is for normative realism, as you rightly said.

Here the argument starts with this thought - What justifies our belief in all sorts of theoretical entities, for instance? Things like electrons or other things of this kind. Often people say, “Well, they are indispensable to our best theories. Try to do physics without quantifying over electrons and tell me how well you’re doing. If you can’t do remotely as well as we can do our physics when we are allowed to quantify over electrons, this we think gives us a good reason to believe that electrons exist.”

There are different ways of making, you know, perfect, precise sense of that. It’s a kind of inference to the best explanation, right? You look at all the phenomena and you push up your best physical theory, you say, that’s my best explanation of all this stuff, and then say, well, if that’s the best explanation, then it’s probably true.

There’s some literature on why this is the case, but for now, I just want you to accept with me that it does make the explanation more likely to be true - that it does make it rational to believe the better explanation. And with it also, in the existence of whatever entities that the explanation quantifies over.

But if you’re going to try and solve the general puzzle, why think that the fact that something is needed for explanation gives any reason to believe that it’s actually true or that these entities exist?

Then I think you’re going to have to resort to what Joshua Schechter and I in other papers call something like indispensable projects or intrinsically indispensable projects. The mere fact that something is indispensable for something else doesn’t give any reason to believe in it unless you’re saying something about the pedigree of that something else to which it’s indispensable. The fact that something is indispensable to witchcraft does not give any reason to believe in that thing, because witchcraft, we think, is debunked. But if something is indispensable for the scientific project of explaining the world around us, that’s respectable enough. Then we’re justified in believing the explanation including in the existence of whatever entities it quantifies over.

If you are with me so far, then the question arises, well, are there any other sufficiently respectable projects? Projects such that indispensability to them can justify ontological commitment, can justify believing the existence of the relevant kind of stuff. I argue that deliberation is one such project. So just as it makes sense to try to explain the world around us, it makes sense also to deliberate, to try and make up our mind about how to proceed, what to do, sometimes maybe what to believe or what to feel.

Adam: What makes a project the kind of project that can ontologically commit us? What is this dividing line between the important projects and the unimportant ones?

What I say in the book is that some projects are rationally required in the following sense: Opting out of them is not a rationally permissible option.

Think about the creature for whom explanation is just never a project. This is not just someone who is not very much into science. That’s totally fine. There are other important things you can do in your life. I mean someone who really never wonders “Why?” about anything in their normal everyday life.

That, we think, is criticizable. So explanation is a rationally non-optional project.

This is significant because usually, there are two rationally permissible responses to finding out that you have a project and that something is indispensable to it. One is to take on board whatever commitments are indispensable for the project and continue with the project. The other is to understand that the commitments that are indispensable for the project are just too much. They are too pricy, they’re sufficiently implausible, and so reject the project, right? Usually both of these are rationally permissible options.

But if there is a project that’s rationally non-optional, just deserting the project is no longer a rationally permissible option. What that means is that there’s only one remaining permissible option, and that’s to undertake the commitments that are indispensable to it. I argue that deliberation is a rationally required project in the same sense as explanation.

Adam: I’m interested in cases where it seems like we do actually have the ability to live life without deliberating. It actually seems rational to do so in some cases. For example, when you’re driving, you won’t be deliberating all the time about how you should be turning the steering wheel, nor should you, because that will be much slower than just allowing your body to do what it does naturally and turn the wheel without deliberating. So it seems like we can opt rationally out of deliberating. Is that a problem? Maybe that’s true and we can live half our lives without deliberating, but in the other half, we have to deliberate. Would that be enough for your argument to go off? And do you think that we have to deliberate sometimes?

David: Yes, it will be more than enough. I don’t think we deliberate half our life. I think we deliberate much less than that, and that’s perfectly okay.

The way I think about the indispensability argument, it says normative truths are indispensable to deliberation and deliberation is sufficiently respectable to justify ontological commitment.

It’s not as if our deliberative needs bring normative truth into existence. That would be an anti-realist view. Similarly, we don’t think that our scientific endeavours bring electrons and the empty set into existence. We think that they reveal to us the fact that those exist independently of our scientific endeavours. So I would say something similar.

The deliberative indispensability argument, if it works, does not tell us anything about how these things are created, it reveals to us the fact that they have already, have always been there. And if that’s true, then really all you need is one deliberation. You just need the total desertion of deliberation to be rationally impermissible. That seems to me like a fairly minimal story.

Adam: So, why should we think that deliberation commits us to belief in normative facts?

David: So, I don’t think that whenever we deliberate, we think explicitly with normative truths in mind. Often we don’t, right? So, I’m trying to decide whether to have dessert or not. It’s always the same for me, namely, it’s tasty, the sugar is going to be really nice right now. I’m trying to lose weight, it’s not very good for me, and you know, the usual kind of stuff. At some point, I suppose I’m going to settle on having the cake now, or not having the cake. I’m not necessarily going to think in terms such as, ah, the best overall decision for me now is so-and-so, that’s a clearly evaluative term, right? Or the weightiest reason is against my having the dessert, that’s a normative claim. Or I ought not have the dessert. I’m just going to consider these different things and decide for the dessert or against it. So, I don’t want to claim that people always use such explicit normative thinking.

What I do want to claim is something more subtle and more implicit, namely that when you are deciding in such cases, you commit yourself to a normative truth. Here’s a test I like to do: think about deliberation in which you decide without weakness of will. Suppose you’re trying to decide between two options, and in this case, you have actually deliberated, and you have fallen on the side of action A rather than B. Suppose then I come to you and say, well, so you think it makes better sense for you to perform action A rather than B?

It seems like you’re going to say, “Sure, I just told you, I thought about it, and I decided to do it.” It’s not going to sound as if I’m revealing to you a new commitment, a thought that hadn’t occurred to you. Maybe you haven’t thought explicitly in normative terms, you haven’t told yourself something like, “Oh, that’s the option that makes sense for me” or “The reasons for A outweigh the reasons for B”. But when I flesh these thoughts out, then you’re going to say, “Yeah, that’s obviously what I thought”, right?

So this will reveal a way in which by bringing the deliberation to its conclusion and deciding what you’re going to do, you have indicated that you already have a normative belief there, something along the lines of “There’s weightier reason to go with this option than that one”.

Adam: The counterexample I was going to give is no longer a counterexample given what you’ve said, because I wasn’t thinking about the end of the deliberation as the relevant fact. I was going to raise the following: when SETI scans space for alien life, we don’t think they’re committing themselves to belief in aliens. You might set up this sensor to look for alien life and think that it’s not going to work because there are no aliens. But I suppose then the relevant analogy now would be them stopping the search and linguistically analyzing a piece of data, which does seem more like them committing themselves to belief in an alien message. In the same way, if you stop deliberating and you take action, it seems like you’re committing yourself to the view that there is really a normative truth here. Right? So this counterexample is no longer any good.

David: I’m not sure. This sounds too flattering to me. So I’m not sure. Let’s try and think it through.

Adam: Well, I had another thought which is -- can someone stop deliberating because they’ve realised that it’s pointless to continue deliberating, because there is no normative truth out there? So, SETI might shut down the system because they haven’t received any alien messages and they think that they never will.

David: Yeah, no, that’s definitely a possibility. First, to deal with this, I think I would like to introduce a distinction between stopping deliberating and bringing deliberation to its conclusion. You can stop deliberating for any number of reasons. You can stop deliberating because it’s lunchtime and it’s time to eat and you can continue deliberating after lunch. But when you stop deliberation in order to have lunch, you’re not bringing deliberation to its conclusion in the intended sense.

I think that deals with that point. And I think you’re right: that wasn’t obvious to me until you said so; in the book I emphasize more the deliberation process, and here in our conversation I emphasize more bringing deliberation to its conclusion.

The reason why you can’t really deliberate without believing in normative truth is that you know that you won’t be able to bring the deliberation to its conclusion without a normative belief, and you can’t deliberate while also thinking that it won’t be possible to bring deliberation to its conclusion. So, I think the two are closely related.

Furthermore, the way the process works is significant. Say you’re asking yourself a question. Should I move to this other town for this other job? You ask things like, what’s the cost of living there? Would the job be more interesting? What about my social life? Whenever you ask these different questions, you ask factual questions, right? But you seem to be presupposing that if the cost of living is sufficiently high, that strongly counts against moving there. That is already a normative judgment. So, it seems to me that a deliberation is impossible for a creature who does not believe in normative truth. We can actually run the following thought experiment. So suppose there’s a creature, maybe an Al kind of creature, I don’t know, like us in many ways, except they don’t have any normative beliefs, or anything like that. Can they deliberate?

I mean, they can certainly do some of the things that are in the vicinity of deliberation. They can fail to choose confidently and immediately. Or they can ask themselves the factual questions that may be the basis of deliberation. But I feel rather confident that you cannot classify them as deliberators. They’re doing other things, but they’re not deliberating.

Adam: Mm-hmm. Because deliberating seems to be just asking yourself, “What should I do?”. Couldn’t our explanation of this simply be, though, that we believe that it’s possible that there’s an answer, rather than that we believe that there really is an answer? With the case of SETI though, doesn’t it seem like it’s more attractive to think the scientists believe that possibly aliens exist rather than that they actually believe aliens exist?

David: That may be so, but if that’s so, that may also be because there’s more data. For instance, what do these scientists say when they ask Congress to fund their project? Some people would say, there are aliens there. Look how humongous the universe is. It just doesn’t make sense that we’re all alone here. I know they’re there. Let us find them.

But most of the scientists are going to say, look, it’s a really interesting question whether there are aliens there. If there are, it would be extremely valuable to know that. If there aren’t, it would be valuable to know that. And so I think our project should be funded. Well, that’s an indication that they don’t really believe it and so on, right? This is just an example of what they’re saying, but of course what we’re interested in is not only what they’re saying.
They may be engaging in the project in a different way. How would they deal with the first 100 failures, right? Well, in one case, they’re disposed to just get more funding, continue to the next 100. And in others, they may be more okay with saying, “Okay, so we can never be certain, of course, this is science after all, but we’re happy to stop at least provisionally and extrapolate. So far we don’t know of any other civilizations. Let’s check again when we have the time and money for it.”

That’s not how we treat normative truth in our deliberations. And we don’t say things like, “Sure, I’m going to ask myself this normative question. If I come up with a normative answer, that would be great. If I don’t, that will in itself be interesting.” That’s not the spirit with which we approach our moral, our practical deliberation.

We just take it for granted that there are answers to our question. We may yet be wrong about this, which is why my book includes all these discussions of objections to realism, in the second half of the book, all these discussions of objections to realism.

So, the difference is that it doesn’t seem like we deliberate and that we’d be interested to find out, actually there were no normative facts. That wouldn’t be a satisfying conclusion to deliberation. So deliberation is different from the alien case in that way.

Subscribe now

Adam: Okay, so supposing this whole argument goes off and we find out that there are normative truths. Couldn’t these be hypothetical imperatives like “Adam should drink his water, but that’s because Adam wants to be hydrated”? This really is a normative truth, but it’s just founded on my desire to be hydrated, and that’s just a fundamental part of me. Couldn’t your argument for deliberative indispensability just bottom out here at hypothetical imperatives rather than categorical imperatives?

David: Here I think I’m echoing a fairly old argument or thought, and that’s the thought that hypothetical stuff derives its normative force from categorical stuff. The mere fact that you want something does not necessarily give you a reason. For that, what you need is for what you desire to be desirable or for you to have a reason to desire it or something along these lines.

If I have really idiotic desires, like I really desire to know the number of leaves of all the trees in my campus combined, that does not give me a reason to, many people will agree, count the leaves, because it’s an idiotic desire. There is no way in which that would be desirable.
People also give the examples of immoral desires, right? If I want to be the most efficient serial killer Israel has ever had, that does not give me a reason to practice my shooting skills, because that’s a horrible desire, there is nothing desirable about it.

It’s only acceptable desires that can give you reasons in this way. They don’t have to be all sacred intellectual desires, but they have to be acceptable. They have to pass some normative filter for them to be able to do this work. So, if that’s true, nothing can stop the normative regress unless it’s categorically normative. Hypothetically normative things - all their normativity derives from something categorical.

Adam: Surely a subjectivist against whom you might be arguing will just deny this and say, “Well, no, actually, it is a normative fact that Adam should drink water if he wants to drink water. And we don’t need to say anything more about that because it for it to be a reason just is for Adam to want it.”

David: That is definitely a way to go, but then they would have to say things about the examples I mentioned. If that’s true, does that mean that when I want to know whether the number of leaves on campus is even or odd, then I do have a reason to count all the leaves? And when I want to become the most efficient serial killer in Israel, is that desire also reason-giving?

Adam: Maybe in these cases, you do have a reason. It’s just that you also have other reasons that outweigh that reason. For example, the fact that it would take forever to count the leaves, and that you want not to waste time.

David: We have such a view in the literature. At least one influential recent author who puts this center stage is Mark Schroeder and he defends his hypotheticalism, which is not exactly the view you mentioned, but it’s fairly close. But we have to worry about the details of these views. So, I want to know more about what you mean by desire. I want to know more about what you’re doing, what you do with these counter examples that I suggested, namely, silly desires and immoral desires. Mark Schroeder has an answer. The answer is partly along your lines, namely, there are reasons, but there are also reasons in the opposite direction. But also, he says, well, reasons come very, very cheap. There are many, many reasons that apply, and for general pragmatist kind of considerations we don’t mention all of them. We only mention the ones that are sufficiently weighty. So now we need an account of the weight of reasons. That’s a whole other story.

Another challenge that comes up for someone like Schroeder is to explain morality. If you think that reasons are in this way somehow reducible to desires, or grounded in desire, or something along these lines, then you have to deal with a person who does not have whatever desires there are that ground moral reasons. Do we think that morality just doesn’t apply to them? That sounds weird. We want it to apply to them and for us to be able to criticize them.

Now, Schroeder is well aware of this and he has discussions in the book in which he tries to fill in the details. I wrote about this book arguing that it fails and the game goes on, but that’s certainly a respectable alternative.

Adam: It seems like another route one of your opponents might take here is just to say “We’ll idealise and say that the person who doesn’t have any desires to be moral is simply not an ideal person. But if they were, if they were perfectly acquainted with morality, and they were in an ideal situation, then they would desire to be moral.” You say in ‘Why Idealize?’ that this sort of idealising is unmotivated. Why is that?

David: The argument in ‘Why idealize?’ is against a host of response-dependent views in meta-ethics. These are views that tie normative truths to something about our responses, for example what we desire - you have a reason to do something if you desire it, or you have a reason to do something if it will partly explain how we desire ourselves to be, or you have a reason to do something if you desire to desire it.

The problem is that when people put forward such response-dependent theories using actual responses of actual people, it’s really easy to come up with counterexamples where you say “Look at that guy over there. He doesn’t desire in any way to look after his kids, and he doesn’t desire to desire it. But surely you agree with me that he has a reason to look after his kids or that he ought to look after his kids, right?” That seems like an initial counter example.

A natural move from the response-dependent theorist, trying to deal with this, would be to idealise: “Well, yes, of course, this guy doesn’t want to look after his kids, but had he been rational, he would have wanted to,” or “Had he enjoyed full imaginative acquaintance with what it’s like to be a deserted child, he would have valued doing so,” or a similar counterfactual idealised condition. This kind of idealisation, it’s important to emphasize, is OK in the context of some theories, such as colour theory. It’s not as if I think idealization is unmotivated across the board in all areas, right? The project is to distinguish between legitimate and illegitimate uses of idealization. In this context, though, of response-dependent theories in meta-ethics, that move always seemed to me like cheating.

In one way it may seem like philosophical progress, because there was a theory which had some plausible stuff going for it, but it also came across counterexamples, so we revised the theory, and now it’s less vulnerable to counterexamples. The problem, though, is that at least sometimes, that’s ad hoc.

It’s ad hoc if the idealization fails to be consistent with whatever it is that led us down the route of this theory to begin with. If what led us down this theory to begin with was something about, you know, surely what’s good for me has to be related to what I care about, then it has to be what I care about and not what I would have cared about under some hypothetical conditions. I mean, maybe it’s a way of avoiding counter examples, but it’s an ad hoc way of avoiding counter examples. It does not have a rationale, nor is it consistent with the rationale for the theory we started with.

Now, idealizing theories come up elsewhere as well. For instance, some people put forward response-dependence theories of color, according to which to be red, roughly, is to be perceived as red. And then we have a similar dialectic: Someone points out cases in which things tend to be seen as red (“to appear redly”) even though they’re not red, or tend not to appear redly even though they are, etc. And there too a natural move is to idealize, and say something like: Well, to be red is not to appear redly, but such as to appear redly to an ideal viewer in ideal lighting. There too, the “Why Idealize” challenge applies – they have to show that this idealization is consistent with whatever led them down the response-dependence path in the first place. And perhaps – I really don’t know – they can. So that the challenge applies quite broadly, but in different contexts different theories may do better or worse in how they respond to it.

So, I don’t argue that all idealization is problematic. What I argue is that the idealization involved in many response-dependent theories in meta- ethics is problematic, because it’s inconsistent with the rationales that push some people down the response dependence line to begin with.

Adam: So, the rationale that I was offering and which you deal with in the paper is to sort of revise the rationale that we began with. Rather than saying, “Well, what’s valuable has to be in line with our desires,” instead we might say, “Well, I don’t know what’s valuable, but let’s just have a look at how people tend to figure out what to treat as valuable.” And the way they tend to figure it out is to imaginatively acquaint themselves with their surroundings and then come to certain conclusions about value. So, a good explanation might then be that that which is valuable is that which someone would desire if they were imaginatively acquainted with it. But you argue there’s a better explanation than this.

David: Right, right. So, the next step is to say -- well, to idealise merely in order to avoid counter examples is not a good rationale for idealization. But maybe there are other rationales. There’s this one you mention, which may be found in the work of several philosophers, including maybe David Lewis. The thought is that the rationale for idealization is not simply cheating and trying to rule out counterexamples. The rationale for idealization is that the standards that I invoke in my idealising are already there in the relevant practice.

For example, you can see that people, when they try, say, to make up their mind about what they value, try to imagine more closely what it’s like to be in a situation where they experience it. The fact that that’s a part of what they do in their normal evaluative lives is an indication that the ideal is an ideal of full information and full imaginative acquaintance. That’s a possible rationale.

What I think is that there’s a simpler explanation for why they are going about their business in this way. The simpler explanation is that they are all intuitively my kind of realists. They know that value is there in the world and they try to view the normative world better.

It’s like, you know, if both of us are waiting for a bus, it’s a distance away and we’re not sure what the bus’s number is, but that’s important to us. Then we may do things in order to improve what we see. I may cover my eyes, or I may put on my better glasses, or you may just wait a little bit so that it comes closer and then you’ll see it right.

The explanation for why we’re doing these things is not that we have some ideal of best visibility. The explanation of that thing is that we know that the number of the buses is independent of us and we’re trying to conduct our epistemic goings-on in a way that increases our chances of getting it right. Similarly, why are people running their intellectual business in the vicinity of value in this way? Not because there’s an independent kind of procedural ideal of how to go about evaluating, but rather because they know or anyway believe that value is there in the world and they’re trying to improve their eyesight vis-a-vis values. So, the realist explanation is clearly better, as it’s much more straightforward. It may yet be that the explanation fails for other reasons, but that’s the natural explanation.

Adam: Yeah, so let’s investigate one of the objections that you try to deal with in the book, which I think is probably the most compelling objection, which is -- if these robust, realist facts about morality and normativity are up there in Plato’s heaven and we’re down here looking up at them, how is it that we come to know what they are? This seems like a problem, because they’re causally inert, so we can’t actually look at them with our eyes or anything. We’ve just got to somehow know that they’re there. The explanation you give is that what’s evolutionarily advantageous is morally good, or at least normatively good. And so our normative beliefs, which have developed through evolution, align in tandem with what’s actually good. Is this an accurate characterization?

David: Yeah. So, there are all these moral truths out there, which if I’m right, don’t depend on us, and the moral facts and entities and properties are causally inert and so on. And then there’s us, and we live in a highly empirical, causal kind of world. And our world includes our moral judgments or our moral beliefs. But wouldn’t it just be a miraculous kind of coincidence if our moral beliefs actually somehow hit upon the moral truth, seeing that there is no relation, certainly no causal relation between them.

In general, if there is a striking correlation between the A facts and the B facts, usually we think that striking correlation is calling for explanation. And that the explanations are typically going to go either by invoking some relations from the A facts to the B facts. So, the A facts cause the B facts, or they partly constitute the B facts or something on the other. Or we’re going to invoke relation in the opposite direction. That’s how we usually think correlations are explained.

But in the case of moral truths on one side, moral beliefs on the other, according to my kind of realism, these kinds of explanations are wrong. We don’t think that the moral facts cause the moral beliefs because we think that the moral facts are causally inert. And we don’t think that the moral beliefs cause the moral facts because that will be an anti-realist view. That will be the view according to which people’s beliefs make the moral truth. So, these natural ways of going are just blocked.

And so, it does seem like a miracle. Notice that if we deny the miracle, namely if we say there is no correlation between the moral facts and our moral truth, then we seem to be committed to suspending judgment on the moral duties. If you had a hundred propositions you believed, P1 through P100, and then God whispered in your ear that at least fifty of them are false, but didn’t tell you which, you would have to really reduce your confidence in all of them and maybe just suspend judgment.

So, it’s really important to continue believing in this correlation, but then it seems like a miracle. We need an explanation for it. And I say, well, it’s a miracle, but it’s a really small miracle, and the way I do this is by noticing that it’s not as if the only way of explaining a correlation between the A’s and the B’s is by something that goes from the A’s to the B’s, or something that goes from the B’s to the A’s.

The further possible explanation is that there’s some C that explains both. It’s kind of a pre-established harmony argument. This is where evolution comes in. I want to say, look, the evolutionary aim, survival or maybe something about genes, is something that deeply matters (as it were) to evolution but it’s also good. It’s usually better for you to survive than to not survive. If we start with that, then I say, you get an explanation of the correlation, not because of the relation between moral truths and moral beliefs, but because both are related to this idea - the fact that survival is good. Our beliefs are partly shaped by the evolutionary aim of survival, and survival is in fact good.

That doesn’t give you, of course, all the correlation you need, but it gives you an initial correlation that then, if you employ general reasoning mechanisms, may become what you need. So, I’m not saying that what matters morally somehow more directly is evolution, or that that’s good.

Adam: So, a problem I see with this is that even if the facts and our beliefs go along in tandem with one another, why would these beliefs about what’s evolutionarily advantageous for us suddenly turn into ideas about morality or normativity? Surely we would just believe “I will survive if I eat food” or “I’ll survive if I don’t murder my friends,” rather than believing “It’s morally good for me to not murder my friends.” Where does the link between evolutionary advantage and morality get made? It seems like these two sets might be extensionally identical, but we don’t have any reason to call the evolutionarily advantageous things moral.

David: You ask, well, isn’t it enough that I just choose the sweeter fruit. Why do I need to also have normative or moral beliefs, that I want to for something like this? And well, that’s a different kind of question. I don’t think I need an answer to it. In general, I would say the following, though, about evolutionary explanations. Evolutionary explanations do not encourage you to go for explanations in terms of optimality.

Some people say such things as “For many purposes, nature would not have granted us reason because instinct would be more efficient. So that means that reason must have another function, right?” That’s a fallacy because nature does not grant us stuff nor does natural selection. It’s highly path dependent, just as if you are here now and this option is better than that one in terms of passing on your genes or whatever, then you are more likely to go down this road than down the other. But that doesn’t mean you’re going to end up with fully optimal kind of arrangements, body, beliefs, and so on. Right? There’s no reason to expect that to happen. So I think that the question, “Well, why do you need those beliefs? Can’t you settle for the mere action?” -- I don’t think that’s how evolution works.

Adam: Okay, and there’s another problem, I think. Would you agree evolutionary aims are not the ultimate good - they can’t account for all the things that are of value? There are some other good things, too. Maybe fairness is also good alongside the evolutionary aims.

It seems like sometimes fairness will override the evolutionary aim and be more important. So it’s true in some cases that we should act against evolutionary aims. Surely your explanation for moral epistemology would allow us never to know that this is the case.

David: Well, my initial story won’t give you all the correlation you need, but it will get you a good starting point. And then you can employ general reasoning mechanisms. So, here’s another thing that’s true of evolutionary explanations in general. It’s quite possible that a mechanism developed because of some evolutionary payoffs or whatever, but then it gives you further options that are not themselves of evolutionary value.

I think that when we’re doing ethics, we’re employing pretty much the same kind of reasoning that we employ elsewhere. It’s not a huge puzzle to think how come we have these abilities, because in many contexts these reasoning abilities are conducive to survival and to procreation and so on.

Maybe good metaphysical reasoning skills are not that evolutionarily valuable, but that doesn’t matter because we’ve acquired generally good and reliable reasoning skills, because of their payoffs when we’re reasoning about tigers or about where to go for food or whatever. We can employ these reasoning skills that we already have elsewhere even when they don’t have evolutionary payoffs. What that means is that if we have an initial state with at least some correlation between our moral beliefs and the truth, we can then subject them to the relevant, purportedly fairly good, truth-conducive reasoning mechanisms, and that will get us closer and closer to the truth.

Maybe I’ll give an example. In other non-moral contexts, we often apply the kind of reasoning norm that says ‘generalise’, ‘reject irrelevant differences’, things like that. Presumably, this is, I’m really speculating here, but presumably having this kind of reasoning norm and skill is actually evolutionarily valuable in contexts that have nothing to do with morality. Because the creatures who have generalized their views, they’ve understood how to light fire much quicker than those who didn’t, or some similar advantage. Now, I believe that my child’s life is of value. And suppose that what partly explains this is an evolutionary story about how it helps me pass along genes on if my child actually makes it to adult age. And then on comes your child.

The question is, should I feel free to murder her? Well, I’ve already acquired these general skills that say, “Wait a second, don’t draw arbitrary distinctions, try to generalize”, and so I will now come to see that actually it’s important that your child gets proper nourishment and not just mine, because there’s no real difference between them.

I thought of this example right now -- I’m not sure it’s sufficiently careful, but it’s supposed to give you a feel for how this is supposed to go. Now, the belief that my child should be nourished, that you could probably explain directly and evolutionarily, that’s not difficult to explain. My belief that your child should be nourished is a harder one to explain evolutionarily. It’s easier to explain it if you’re allowed to make use of these reasoning mechanisms that have presumably evolved in virtue of their utility elsewhere.

Adam: That’s very interesting. I’ll have to think about that example some more and see if it helps, but we’re going to run out of time now. Thank you very much for speaking with me - this conversation has been really good.

David: Thank you; it’s been a pleasure.

Subscribe now

In this play we see a philosophical showdown involving the following characters:

• The Atheist (A)

• The More Skeptical Theist (MST)

• The Less Skeptical Theist (LST)

• The Minimally Skeptical Theist (MiST)

• The Contextualist Skeptical Theist (CST)

As we shall see, the Atheist sees off the arguments of MST and LST, successfully showing that their accounts of skeptical theism lead to general skepticism. MiST and CST however, do not, and have good accounts of where the problem of evil go wrong, which do not in turn lead them into general skepticism, despite the atheist’s attempts to show that they might and to revive the problem of evil.

Pictured: The ancient theatre in which this play is performed.

Act I: In which the Atheist defeats the More Skeptical Theist

Atheist (A): So, you think there’s a tri-omni God? Well, I reckon I have some strong evidence against that idea – just look around you! This God of yours has allowed all sorts of evils to occur: animals being torn to shreds by other animals, humans being systematically murdered, and everyday annoyances like stubbed toes and hangnails. I can’t see any reasons why a tri-omni God would allow these evils, so I can confirm that He wouldn’t. And since a tri-omni God wouldn’t allow such things, and we can see them occurring, we should assume the reason He doesn’t stop them is that He doesn’t exist!

More Skeptical Theist (MST): So you think there are some evils out there that an all-good God wouldn’t permit?

A: Exactly.

MST: Okay, let’s call them gratuitous evils.

A: Why do we need to name them? This conversation is over; your God does not exist.

MST: Well, hang on a second. I think you’ve made a mistake in your argument.

A: What’s that then?

MST: I don’t think you know that there are gratuitous evils!

A: Why not?

MST: Well, in order to know that an evil is gratuitous, you would need to know about divine psychology, understanding God’s reasons for allowing certain things to occur or for causing certain things. But you don’t have a good reason for believing that you do understand God’s psychology!

A: So what?

MST: If you don’t have good reason to believe you understand something then you don’t know things about it! If you don’t have good reason to believe you understand quantum mechanics, then obviously you don’t know that the Copenhagen interpretation is false.

A: Maybe that’s true about quantum mechanics, but it’s obviously not true for everything.

MST: No? Surely, for you to have knowledge of x, you should have good reason to believe you can judge matters such as x.

A: Most definitely not! Let’s plug in some propositions to your rule and test it: let x be the proposition ‘There is a tree’.

Pictured: A tree! Or is it…?

MST: Sure, that’s fine! For me to have knowledge that there’s a tree, I need good reason to believe I can judge matters such as there being a tree. I do have good reason to believe this! I have eyes, there’s no mist –

Minimally Skeptical Theist (MiST): Did someone say my name?

MST: No, shut up. As I was saying: I know there’s no mist and that my eyes are working in the right environment to see whether or not there’s a tree, so I have good reasons to believe that I can judge whether there’s a tree. So, I know there’s a tree!

A: Aha, but how do you know there’s no mist and that your eyes are in the right environment?

MST: Well, they’ve worked for me for the last few decades.

A: And how do you know that?

MST: I don’t need to know that to trust my eyes. Gosh, you just sound like a general skeptic!

A: I’m not a general skeptic. I’m just applying your rules! In order to know there’s a tree, we need to be able to give good reasons for trusting such judgements, and to have good reasons for trusting those judgements, we need more reasons for trusting them, and so on.

MST: Okay, maybe you have a point and my rule has some worrying consequences. But what do you propose as an alternative?

A: In some circumstances, you don’t need to give reasons for believing in what your faculties tell you. I can’t always give good reasons to believe my eyes are reliable, but I can still trust them. I can’t always give good reasons to believe my doctor isn’t lying to me, but I can still trust her. There may well be good reasons, but we don’t always need to be able to give them. That’s my rule. If we were to follow your rule instead and require good reasons all the time, I would very quickly fall into general skepticism, believing that I don’t know anything!

MST: Okay, so I guess requiring good reasons to believe we can judge God’s reasons was a step too far. Ah, but here comes my brother, the Less Skeptical Theist! I’m sure he’d like a word with you.

Act II: In which the Atheist defeats the Less Skeptical Theist

Less Skeptical Theist (LST): Hello, Mr. Atheist!

A: Hello! I hope you don’t agree with your brother’s silly rule.

LST: I don’t! But I do agree with him that you can’t know that there are gratuitous evils. And I have a rationale for you that doesn’t fall into skepticism like my brother’s does.

A: Ah, let’s hear it then!

LST: Okay, here it is: you can’t know that there are gratuitous evils, because you don’t know about God’s reasons for allowing certain things to occur.

A: It sounds like you’re just the same as your brother!

LST: Well, I’m not! The difference is, while MST said that you don’t know God’s reasons because you couldn’t give him reasons to believe you do, I say that I can offer you my own reasons to believe you don’t know about God’s reasons!

A: Ah, well what might they be?

LST: Well, here’s one: we don’t know what goodness looks like!

A: Oh dear, that’s awfully silly too!

LST: Really? Why?

A: Did you even read ‘Everything Evil II: Depravity and Design’?

LST: Umm… no. Why?

A: As discussed in that post, if you want to continue making the Fine Tuning Argument – hang on, you do want to make the Fine Tuning Argument, right?

LST: Yes, it’s possibly the most scientific and telling argument in favour of the existence of God!

A: Well, in that case, you ought not argue that we don’t know what goods God might aim at, because if you do, you undermine the Fine Tuning Argument. Instead you should say we don’t know what methods God might use to achieve His goods.

LST: Fine, I’ll do that! I have arguments about methods too! Don’t be such a nit-picker!

A: Okay, let’s hear the reasons why I don’t know what methods God might use to achieve His goods.

LST: Okay, well here’s one very good reason: we don’t know what the connections there are between goods and evils and we don’t know what goods and evils in the future might be, so God might be able to use a method involving evils in order to bring about goods or prevent evils in the future, and we would be none the wiser!

A: Alright then, sounds like you have a good reason!

LST: Let’s go! I knew you’d come round to my view. I also have more reasons, and there are many in the literature, but we needn’t get into that. I’ll be glad to welcome you at church on Sunday.

A: Not quite. I still have a problem. I think your view leads to general skepticism too.

LST: What! Why?

A: Well, you said I can’t know about God’s reasons for allowing certain things, or, put differently, I can’t know what methods God might use, right?

LST: Right.

A: That means that I can’t know God won’t deceive me about the external world, the past, etc. etc.!

LST: Oh crap.

A: And if I can’t know God won’t deceive me - that is: I can’t know that one of His methods for bringing about good wouldn’t be deceiving me - then I don’t know facts about the external world, the past, etc. I could be being deceived!

LST: Darn it, you’re right! Well, good thing my other brother, the Minimally Skeptical Theist, just got here!

Act III: In which the Minimally Skeptical Theist avenges his brothers

Minimally Skeptical Theist (MiST, who wears sunglasses indoors and is ten times cooler than his brothers): Oh, brother LST, you are very silly! Your reason for doubting knowledge of God’s reasons is not what it claims to be. It doesn’t stop us knowing about God’s reasons at all!

LST: Doesn’t it? Then how can it stop Mr. Atheist’s argument?

MiST: Let’s rewind the tape.

— A fancy transition plays as we cast our minds back to Act I —

A (in flashback): This God of yours has allowed all sorts of evils to occur: animals being torn to shreds by other animals, humans being systematically murdered, and everyday annoyances like stubbed toes and hangnails. I can’t see any reasons why a tri-omni God would allow these evils, so I can confirm that He wouldn’t.

— Another fancy transition plays as we come back to the present —

MiST: See that inference the Atheist just made? He can’t see reasons, so there aren’t any? That’s what philosophers call a noseeum inference.

LST: Why do philosophers have to use silly names for everything?

MiST: It makes them feel clever. Anyway, a noseeum inference basically says: I can’t see ‘um, so they ain’t there.1

A: So, what? Noseeum inferences can work! If somebody asks me whether there are elephants in my garage, I can make a noseeum inference about it – I can’t see any elephants in my garage, so they ain’t there!2

MiST: Ah, but what my brother’s point about connections does show us is that can’t make a noseeum inference about God’s reasons. God’s methods and the goods and evils possible in the future are so far outside of our understanding that noseeum inferences about God’s reasons are invalid.

A: Okay, MiST, so maybe you’re right, and there’s a reason to doubt the argument from evil. But how are you going to avoid skepticism?

MiST: Why would skepticism be a problem for me?

A: Well, if you can’t make a noseeum inference to conclusions about God’s reasons, then you can’t know God wouldn’t deceive you! You might sit there and think ‘Well, I can’t see any reason that God might deceive me, so He wouldn’t’, but as you’ve just said, that kind of reasoning is invalid!

MiST: That’s not how I conclude that God wouldn’t deceive me.

A: No? Then how do you form such a conclusion?

MiST: Have you ever heard of the externalist response to skepticism?

A: No.

MiST: Perfect, because I was going to tell you about it anyway. Basically, externalists (of which I am one) say that so long as we are not actually in a skeptical scenario like being in the matrix or being deceived by God, we know that we aren’t.

A: And how do we know that?

MiST: Well, we have externalist justification for our beliefs like ‘I have hands’, and –

A: Hang on, what’s externalist justification?

MiST: Ah well, that just means you have knowledge justified by processes you don’t necessarily have phenomenal access to. For example, the fact that your eyes are reliable isn’t necessarily a fact which you can access phenomenally, but it still justifies your beliefs formed with your eyes.

A: Ah, well that’s exactly what I was saying to your brother, MST, earlier!

MST: You don’t need to remind me!

MiST: So, anyway, we have knowledge through externalist means of beliefs like ‘I have hands’, and this knowledge rules out the skeptical possibilities; the knowledge that I have hands entails that I’m not a brain in a vat, or in the matrix, or being deceived by God. This is why I told LST that his point doesn’t mean we can’t know about God’s reasons. In fact, we can know about God’s reasons via this externalist argument!

Pictured: A trusty old brain in a vat

A: Ah, but I have read about this externalist response, and I don’t find it convincing.

MiST: Why not?

A: Well, Stephen Law argues that the situation you’ve put yourself in by being a skeptical theist undermines even externalist knowledge.3

MiST: How’s that then?

A: Law uses this example: Imagine you’re sitting at your desk looking at what appears to be an orange. Now imagine, further, that you know there’s a man called Olly who has a special projector that can project a hologram of an orange, and that he has rolled a die to decide whether he will in fact project that hologram onto your desk. It seems like even if you’re an externalist, you should not consider yourself to have knowledge that there’s an orange in this situation, even if there actually is an orange.

MiST: Maybe you’re right about that. My preferred version of externalism, proper functionalism, says that your eyes need to be working in an appropriate environment in order to confer epistemic justification on your beliefs formed through sight. But an environment in which whether I’ll see an orange depends on the roll of a die is not an appropriate environment!

A: Right! And that’s exactly the sort of situation we’re in with God.

MiST: Hang on! How did you figure that?

A: Well, we don’t know about God’s reasons!

MiST: That doesn’t mean the question of my deception is being determined randomly! It’s entirely determined by the facts about whether that would be the right thing for God to do. To make the ‘Olly’s Orange’ scenario analogous, what we should say is that I can’t noseeum inference my way to conclusions about Olly’s decision about the hologram.4

A: How does that help?

MiST: Well, now we clearly can have knowledge that there’s an orange there. If there is in fact an orange there, then my eyes are working properly, and so I know Olly isn’t using his hologram projector.

Pictured: Olly’s orange

A: Well that’s not a convincing answer to the thought experiment!

MiST: Take that up with externalists, not with skeptical theists. That’s more their department. Moreover, I can have other ways of knowing that Olly isn’t using his projector, still without making noseeum inferences. He could just tell me that he isn’t using it!

A: Well, that doesn’t work for God!

MiST: Why not?

A: Because skeptical theism also entails that we shouldn’t trust divine testimony!

MiST: Okay, well let’s discuss that another time! When Adam has written about that, he’ll come back here and add a link giving his verdict on that question. But even if you’re right, there are other ways to form conclusions about Olly’s reasons, and by extension God’s reasons, that still aren’t using noseeum inferences. As Perry Hendricks argues, we can use Alstonian epistemology, or phenomenal conservatism to do so.5

A: Fine! Fine. So let’s say the skeptical theist can use some method to come to conclusions about God’s reasons. It sounds like your favourite route is externalism – if we aren’t being deceived then we know that God doesn’t have reasons to deceive us, because knowledge like the knowledge that I have hands rules it out.

MiST: Exactly. And we may not know that we have knowledge that God isn’t deceiving us, but we do in fact have this knowledge, just as long as we’re not actually being deceived.

Share

A: Okay, that sounds reasonable. But now you’ve given me another idea…

Act IV: In which the Atheist uses other problems of evil

A: You may have defeated the noseeum version of the problem of evil, but --

MiST: Wait, there are multiple versions of the problem of evil?

A: Yes, didn’t you read ‘Everything Evil I: Plenty of Problems’?

MiST: Umm… no.

A: Ok, well now that you’ve mentioned externalism, which justifies a commonsense solution to skeptical worries, you’ve reminded me of the commonsense problem of evil!

MiST: Uh oh.

A: The argument goes like this: maybe I can’t make noseeum inferences to conclusions about God’s reasons, but can’t I make commonsense inferences? I would say something like: this evil really appears to me to be gratuitous, and I don’t have any defeaters for this belief, so I can conclude that it is gratuitous.

MiST: What!

A: What?

MiST: You really have seemings (or appearances, or intuitions, or whatever you want to call them) about whether an evil is gratuitous or not? Without looking at the goods and evils it might cause? That’s like looking at Everest and saying it just appears to you to be the tallest mountain, without looking at any other mountains. Surely, you don’t have seemings like that!6

A: Well, I do!

MiST: Huh, that’s odd. Okay. I’ll suppose for the sake of argument that you’re telling the truth. Even so, I think there are actually defeaters for these seemings.

A: Like what?

MiST: Well, when we really put our heads down and try to calculate whether the consequences of certain things make them good or evil, we have a really hard time. It’s hard to see the connections between different events, and it’s hard to properly evaluate those events. So, even if we’re trying really hard it can go wrong judging whether something is an evil. So why would we trust our seemings if our calculations are unreliable?7

A: Well, I think actually we can do such calculations. And if I’m right, this means that the commonsense problem goes through as well as another: the calculation problem! All I need to do is do one good calculation of the consequences of an event, and this will not only legitimise the calculation problem, but also perhaps get rid of the defeater you just proposed.

MiST: Well, no, you would need to do many calculations and somehow verify them in order to show that your calculations are reliable. But let’s just say you have to do one, to which I say: Okay, do a calculation, then!

A: Ummm. Well… ummm. Some other time!

MiST: Right, well if you won’t display your excellent calculation powers about gratuity, I guess we’re just at an impasse.

A: Right we are.

MiST: Well, until you calculate for me, I’ll remain happy in my skeptical theism!

A: And… well… I’ll remain happy with my atheism, because I know I can calculate these sorts of things… …right?

Act V: In which the Contextualist Skeptical Theist gets a turn

Contextualist8 Skeptical Theist (CST): Wait guys, you can’t finish the conversation without me! I didn’t even get a turn!

MiST: Why would we let you have a turn? I did very well on my own.

CST: Ah, but externalism is not the best way to respond to skepticism! Contextualism is! Also, do you not see the name of the blog? If an account dedicated to David Lewis didn’t discuss contextualism, then that would be the real gratuitous evil.

Subscribe now

MiST: Alright then, have your turn, I’m out of here.

A: Okay, so you, CST, think that externalism is silly! Surely, then, you should agree with me that God could be deceiving us and a skeptical theist doesn’t have everyday knowledge.

CST: Right!

A: Wait, really?

CST: Well, not really.

A: Oh.

CST: You’re right that when the skeptical theist is considering the possibility of God deceiving them, they don’t have knowledge of everyday propositions like ‘I have hands’. This is because they can’t rule out the possibility of God deceiving them. But in their everyday lives, they do have knowledge of ‘I have hands’ because they rule out the deceptive God possibility by convention. And they are right to do so.

A: Wait, that’s crazy! This contextualism sounds weird and you’re just bringing it up to defend theism!

CST: Nuh-uh! Contextualism is actually motivated by other considerations regardless of theism. Here’s one example: Imagine you’re looking at a donkey in a zoo.9 You know that it’s a donkey. But then somebody comes along and says ‘What if it’s a cleverly disguised mule?’ You realise that you can’t rule that possibility out because you don’t know enough to distinguish a real donkey from a cleverly disguised mule, so you conclude you don’t know. But it seems that you really did know it was a donkey before the smart-alec intervened. Contextualism saves this intuition by saying that the introduction of a previously ruled-out possibility (even if before it was just ruled out by convention) undermines your knowledge. You went from knowing there was a donkey, to not knowing.

Pictured: A donkey. Or is it…?

A: Okay, well I –

CST: Wait! There’s more! Here’s another example: Sue is driving home from holiday and wants to cash a cheque.10 She thinks about whether the bank will be open when she gets back and thinks about the last time she went there on a Saturday. It was open on Saturday a few months ago, so she concludes it will be open when she gets home tonight. It seems in this situation Sue has knowledge. However, now imagine that she needs to cash the cheque tonight or it will bounce and she won’t get any of the money. Suddenly, it seems like Sue no longer has knowledge. This is called stakes-sensitivity, which means that we can lose knowledge if the stakes are made higher and the evidence remains the same.

A: So, what?

CST: Well, this is informative regarding the problem of evil. On the question of whether God exists, the stakes are really high, because if you answer the question wrong you might end up going to hell or wasting your life pursuing a relationship with someone that doesn’t actually exist. So, when we are trying to determine whether He exists, the stakes are high and so the standards for knowledge are high. This means we can’t rule out certain possibilities like there being a God-justifying reason for an evil occurring. However, if we think that God exists, when we go about our daily lives the standards for knowledge are not so high as to leave relevant the possibility that he might be deceiving us. In fact, because God wants the best for us, if He is deceiving us it’s for our own good, so the stakes are low on that question. Therefore, we can forget about the possibility that He’s deceiving us and still say we know that we have hands, for example.

A: Ah, but you’ve made two mistakes!

CST: And what are those?

A: Well, first of all, if it’s like your donkey example or your bank example then there’s a problem. In those examples, we were right to rule out certain skeptical possibilities until somebody introduced those possibilities to us (as in the donkey case) or until the stakes got high (as in the bank case). So, when you say that we can’t rule out the possibility of God deceiving us when we’re considering the problem of evil because the stakes are high, doesn’t that mean that when we’re not considering the problem of evil, the stakes are lower and we can rule those possibilities out? That would mean that in everyday situations we do in fact know that God wouldn’t allow the evils we see and so that He doesn’t exist, and it’s only when we consider the problem of evil that we cease to know He doesn’t exist.

CST: NO!

A: Well, why doesn’t it work like that?

CST: Because the God case is not a straightforward analogy to the donkey case or the bank case. I didn’t mean to say that the problem of evil is an instance like those in that we know our proposition when the stakes are low and we don’t know it when the stakes are high. I think the point LST gave all the way back in Act II shows that we don’t have knowledge, even before contextualism gets involved. I was just adding on to that point by adding that the stakes are high when we consider the problem of evil, so this is even more reason to think that we don’t know. Even when we’re not considering the problem of evil, the defeater given by LST still stands, and so we don’t have knowledge of gratuity.

A: Aha, okay that makes sense. But there’s still the second mistake!

CST: And what’s that?

A: Well you said that if God is deceiving us, it’s for our own good, so the stakes are low. But you forgot about a certain possibility: what if God is deceiving us not for our own good, but for the good of all the universe and at our expense, perhaps dooming us to hell or some other awful fate, for the sake of benefitting the rest of the universe through our deception. This is surely a high stakes situation, making the standards for knowledge greater, and therefore destroying our knowledge that God isn’t deceiving us.

CST: Ah, I suppose that’s a good point. I was wrong earlier when I said it would be low stakes. Sometimes it will be high stakes.

A: Aha, so I’ve got you!

CST: Not quite.

A: Darn it!

CST: I think there are still reasons to believe God wouldn’t do such a thing. Well, so long as you’re not a consequentialist. If you’re a consequentialist, you should think it is actually acceptable for God to deceive you at your own expense if the good it does for the rest of the universe outweighs the harm it does to you. So sorry if you’re a consequentialist; I can’t help you11. But according to lots of rival moral theories, this would be unacceptable for God to do because any harm done to one person must result in greater goods for that person. If you’re not a consequentialist, then, we can rule out the possibility that God might deceive us in this manner.

A: Okay, that makes sense, but aren’t you appealing to moral knowledge?

CST: Yes, and…?

A: Well, I think that skeptical theism leads you to doubt moral knowledge, too!

CST: Well, like we did earlier with divine testimony, let’s leave that aside for another blogpost. I’m sure Adam will add a link to that blogpost when he gets round to writing it.

A: Okay, let’s grant for the sake of the argument, then, that skeptical theism doesn’t lead you to doubt moral knowledge in particular. However, you’re still making an appeal to moral knowledge, and that isn’t allowed, because this is a debate about whether skeptical theism leads to general skepticism! Moral knowledge is included in that skepticism, so you can’t rely on it.

CST: Well, I have two points to make here. Firstly, the challenge was initially framed as leading to skepticism about the external world, the past, etc. Moral knowledge is (arguably) a priori and therefore (also arguably) not included in the gamut of things that we should doubt because of skeptical theism. Secondly, even if it is actually included in that gamut, we can still make appeal to it to justify our worldview because we do not grant that skepticism is true.

A: So you’re going to make a circular argument?

CST: All arguments about how we know things should be in some sense circular, because if they didn’t support themselves they wouldn’t be very reliable theories of knowledge. Furthermore, externalism and coherentism offer some convincing accounts of how theories surrounding knowledge can be circular but still convincing.

A: Ah, so you’re not really a contextualist, but an externalist! Your brother, MiST, was on the money.

CST: Maybe externalism and contextualism need not be enemies…? Anyway, note the general pattern that whatever argument you think should apply to defend against skepticism in the normal situation outside of debates about theism can also be applied by the skeptical theist to defend when the atheist says that skeptical theism leads to general skepticism. There are other responses to skepticism which could be adopted by a skeptical theist in this way12.

Conclusion

To conclude, then, some versions of skeptical theism, namely MST and LST, do lead to general skepticism, but this is of no worry to the theist, as they have the more conservative and well-defended views of MiST and perhaps (if they like contextualism) CST to help them. The skeptical theist, therefore, need not fall into general skepticism. Next time, as promised, I will deal with the objection to skeptical theism that it leads to moral skepticism. At some point in the future we will also deal with the challenge to divine testimony.

Empiricism is the view that all of our justified beliefs ultimately depend on experience. This means that many trivial beliefs are justified, such as the view that you are reading a piece on substack, or that you are looking at pixels. You have experience of these things, so they are justified. However, the view, in the hands of many philosophers, goes further and says that things which aren’t justified by sense data are also justified experientially. For example, when you read Edmund Gettier’s famous paper on justification, you think about the thought experiments he gives and you form the belief that the beliefs considered do not constitute knowledge. This belief, empiricists say, is also justified by experience - the experience of considering the thought experiment and our intuitions and understanding of terms like ‘knowledge’. Thus, empiricism is equipped to affirm the work of lots of philosophy, by recasting what seems like a priori reasoning as actually experiential reasoning. It may make some enemies in the process, particularly rationalists and other defenders of the a priori, but this can be expected of any view that describes how knowledge is acquired.

However, empiricism, which is perfectly respectable by itself, is often confused with, or lumped together with another view called epistemic verificationism. Epistemic verificationism is the bastard son of empiricism and semantic verificationism. During my time at the Oxford Socratic Society, a philosophy society for undergraduates, I often heard people declaring themselves empiricists, and then arguing along the lines of epistemic verificationism, thereby tarnishing empiricism’s good name. While empiricism takes the scrawlings of philosophy and gently corrects its spelling, epistemic verificationism throws philosophy’s work in the fire and sends philosophy to bed with no dinner.

So what is epistemic verificationism? It is the view that only beliefs which can be verified using sense-data are justified. Thus, anyone who particularly values scientific methods and reasoning might be tempted to adopt such a view. In the hands of my friends at the Socratic Society, the view is used to dismiss entire fields of philosophy: moral beliefs aren’t verified using sense-data, so they are unjustified - there is no such thing as morality; much of modern metaphysics is done from the armchair so it must be mistaken; religion consists of beliefs about God which can’t be tested with my eyes and ears, so these beliefs are unjustified1.

Much of this talk goes about under the label of empiricism, garnering the ideas lots of gravitas from the notion that they were held by heavyweight empiricists like Francis Bacon and David Hume. However, as I have hopefully made clear, they are not at all the same view - while empiricism can in principle make room for morality, metaphysics, and religion, epistemic verificationism cannot.

Furthermore, epistemic verificationism is self-refuting while empiricism is not. Empiricists can affirm the principle of empiricism by engaging in thought experiments about justification and realising that all forms of justification are based on experiences. Thus, empiricism can affirm belief in empiricism to be justified.

Epistemic verificationism on the other hand, cannot affirm that belief in epistemic verificationism is justified. That is because beliefs about justification are not verifiable by sense-data. (Belief that another belief is justified comes about through intuition and reflection, not from using your five senses to pick up traces of ‘justified-ness’ in the air around people who have justified beliefs.) Therefore, epistemic verificationism should be rejected because it tells us that belief in it is not justified.

A close friend (or parent) of epistemic verificationism, which I mentioned earlier, is semantic verificationism. Semantic verificationism is more respectable than epistemic verificationism, because it does not refute itself: the view says that to be meaningful a statement must be either analytic or verifiable, and its proponents can argue that the view itself is analytically true, so it does not undermine itself despite not being verifiable. However, I want to borrow from a defence of semantic verificationism, made by the ever-insightful here. Though it may not be necessary to defend semantic verificationism, it may be necessary to defend epistemic verificationism, so we should give it a fair shot.

Carnap’s defence of semantic verificationism says, to quote Silas:

Internal questions are questions of whether or not something is true according to a particular linguistic framework. So we might ask whether numbers exist. And would you look at that: According to the language of Peano arithmetic they do, since we positively quantify over them!

External questions, however, are questions of whether or not something is “really” the case. Do numbers really exist? That’s just a nonsense question, according to Carnap. You haven’t specified a precise interpretation of the question in a language, but are asking outside any linguistic framework—and as such your question is meaningless.

The best we can do for external questions is give pragmatic answers as to which linguistic framework we ought to use—do we want a framework that commits to numbers or not? But if you choose a framework that does commit to numbers, that doesn’t show that numbers really exist.

Likewise for meaning: Whether or not a statement is meaningful is either a question with an analytic answer (when internal) or a non-cognitive pragmatic question about the design of a language (when external).

Carnap goes on to say that the question of whether semantic verificationism should be adopted is an external question, and so should be answered pragmatically. We don’t need to consider whether semantic verificationism is true, we just need to consider whether it is helpful as a rule for our language going forward.

How might this argument be changed up to work for epistemic verificationism? I take it that such an argument would go something like this:

Epistemic verificationism (something must be verifiable by sense-data in order to be justified) is not a statement to be believed or shown as justified, but instead a rule to be followed in good practice of reasoning. We do not need to justify it, just to give pragmatic reasons to adopt it.

However, I do not find such a response convincing. When people give pragmatic reasons to adopt a rule, they are often (and perhaps always) implying truth-claims or justification-claims. When I tell a friend not to dig straight down in Minecraft, and I give the pragmatic reasoning for adopting such a rule (you might fall into lava), I am implying the hidden truth-claim that it is bad to dig straight down, and the justification-claim that the belief that it is bad to dig straight down is justified by the possibility of falling into lava when you do so. Similarly, if I tell a friend they shouldn’t trust any complex judgement-calls they make while drunk, and I give pragmatic reasons for such a rule, like ‘You don’t think clearly when you’re drunk’, I am implying a justification-claim like ‘Complex judgement-calls made while drunk are not justified’. One cannot separate the pragmatic reasons from the truth- and justification-claims.

What does this mean for epistemic verificationism? Well, in response to the Carnapian defence which I proposed above, I am arguing that I simply don’t understand how to pragmatically argue for a rule without making claims about justification and truth. Thus, it is impossible to give pragmatic reasons for adopting epistemic verificationism as a rule if we are not also saying that epistemic verificationism is true or that belief in it is justified.

To explain this further, the Carnapian argument says that an epistemic framework is chosen on pragmatic grounds like simplicity, fruitfulness, etc., but these seem like epistemic virtues - Carnap’s appeals to them can only be understood as epistemic claims that these things lead to knowledge. Carnap is unable to escape the epistemic norms of a framework and evaluate that framework without appealing to those very same norms. It is impossible to justify an epistemic framework without using one.

To conclude, empiricism should not be confused with epistemic verificationism, which is self-refuting. Though Carnap’s argument defending semantic verificationism can be adapted to defend epistemic verificationism, it ultimately fails.

Subscribe now

This is, then (unless I am far less convincing than I believe, or unless is far less well read than I realise), the first response of its kind to the NBLO - the Non-Biological Life Objection, as I have decided to call it.

Here is the Fine Tuning Argument, as laid out by Epsilon:

1. The likelihood that our universe is fine-tuned for biological life, on naturalism, is very small.
2. The likelihood that our universe is fine-tuned for biological life, on theism, is not very small.
3. Thus, our universe being fine-tuned for biological life strongly supports theism over naturalism.

As you may be able to guess from the name, the Non-Biological Life Objection to the Fine Tuning Argument (hereafter ‘NBLO’ and ‘FTA’), posits that God could create lots of life that is completely unlike the life we are familiar with. This idea is, in Epsilon’s post, inspired by the Noroi (呪い) from Jujutsu Kaisen, but in other philosophical works, such as Sinhababu’s famous paper, is displayed with the idea of electrons and protons that are given souls by God and fall in love with each other. If Epsilon and Sinhababu are correct, then, God could create one of any number of possible universes tuned for other kinds of life, be it the spirits from anime shows, or rather ensouled electrons. This would make Premise 2 in the FTA now difficult to defend. If God could create any kind of life from an infinite set of kinds of life, why would we expect him to make biological life? We would not, Epsilon thinks, and so Premise 2 is false and the FTA does not run.

Now, it is important to note what exactly is going on here. Standard proponents of the FTA take it for granted that most of the standard model laid out by physicists is already in some way fixed into place when God comes along and gets involved, messing about with constants such as the cosmological constant. I am yet to read a piece that explains clearly why FTA-proponents make this assumption, but I think those proponents are a lot smarter than I am, so I would not be surprised if there were a good reason. Nonetheless, Epsilon wants to drop this assumption, by pointing out that God is ‘maximally powerful’ and that we should therefore consider Him capable of changing the standard model (or a subset of the laws within it) in order to allow for non-biological life. I think the FTA will run perfectly fine with or without this assumption, so let us, for the sake of entertaining Epsilon’s argument, drop it.

Now, with all the different laws considered as contingent, meaning God can change them and that there are possible worlds in which they are all different, it seems to me that Epsilon is correct and Premise 2 is false. There are infinitely many possible universes that contain life, only a tiny slither of which contain biological life.

However, it still seems that there are a vastly greater number of possible universes with no life than there are with life. Firstly, for every universe with a standard model like our own and a cosmological constant finely tuned for life, there are googols of universes with a standard model like our own and a cosmological constant that does not permit life1. Yes, as Sinhababu says, God could create life in any of these universes by ensouling electrons and protons, but this would require a very particular set of psycho-physical laws, meaning that for every such universe with electron and proton life, there are googols of universes with identical physics but the wrong psycho-physical laws to allow electron and proton life to come about. Yes, as Epsilon says, for every biological-life universe there are many universes with non-biological life, but similarly, for every such universe there are googols of similar universes with the wrong physical laws and/or the wrong psycho-physical laws to allow the kind of life Epsilon is imagining.

Thus, it seems that any universe which contains life is still surprising and well-explained by theism, so the theist can respond to the NBLO by tweaking the FTA. The new one should look something like this:

1. The likelihood that our universe is fine-tuned for biological life, on naturalism, is very small.
2. The likelihood that our universe is fine-tuned for biological life, on theism, is not very small.
3. Thus, our universe being fine-tuned for biological life strongly supports theism over naturalism.

The normal FTA [left] and the suped-up amended FTA with the word ‘biological’ crossed out [right].

That does it for my response to the main thrust of Epsilon’s argument. Casual readers, leave here. Now, for the dedicated, I will respond to the rest of Epsilon’s article, which I think gets some of the dialectic about the FTA wrong:

An extra note on responses to the FTA

In Epsilon’s article, he says that the theist might hope to respond to his NBLO by saying something like:

it is rather expected that there would be some world fine-tuned for biological life. After all, an all-loving God would want to share his love with all sorts of entities. As a result, he would try to maximize the variety of life; so, eventually a world like ours must pop into existence.

And this argument (lets call it the ‘Eventually Response’), Epsilon says, fails because it commits the inverse Gambler’s fallacy.

What is the inverse Gambler’s fallacy?

To understand that we need to back up a bit, and look at the original FTA:

1. The likelihood that our universe is fine-tuned for biological life, on naturalism, is very small.
2. The likelihood that our universe is fine-tuned for biological life, on theism, is not very small.
3. Thus, our universe being fine-tuned for biological life strongly supports theism over naturalism.

In the literature, there is an objection to the FTA, which we will call the ‘Multiverse Objection’. It says that Premise 1 is false, because there may well be (or in some versions there is), a multiverse made up of many universes, each of which have different physical laws. This would make Premise 1 false, because with many universes, it is far less surprising to find one that has the right physical laws to allow biological life to come about. Another way of putting the Multiverse Objection is by saying that the FTA doesn’t really give evidence for theism, it gives evidence for the existence of a Multiverse.

However, the Multiverse Objection, according to many theists and according to Epsilon, commits the inverse Gambler’s fallacy. Finally! We are here, and can explain what the inverse Gambler’s fallacy is. Theists (and Epsilon) who argue that the Multiverse Objection commits the inverse Gambler’s fallacy say that the existence of a multiverse would only explain the fact that some universe had finely-tuned laws, but would not explain the fact that this universe had finely-tuned laws. What does this have to do with gambling, you ask? As Epsilon quotes, so shall I - Philip Goff says the following:

Let’s say you and I walk into a casino,and the first thing we see is somebody having an extraordinary run of luck, rolling double 6 after double 6. I say, ‘Wow, the casino must be full tonight.’ Confused, you ask me to clarify my line of reasoning. I say, ‘Well, if there are only a few people playing in the casino tonight, it’s improbable that you’d get somebody having such an incredible run of luck. But if there are very many people playing, it’s not so improbable that one of them would happen to have a run of good luck.’ This reasoning is also fallacious. We’ve only observed one person, and how many other people there are in the casino has no bearing on the odds of the one person we’ve observed rolling well. This is the inverse gambler’s fallacy. (Emphasis Epsilon’s)

I’ll lay my cards on the table now (pun very much intended) and declare that I don’t believe the Multiverse Objection commits the inverse Gambler’s fallacy. Goff’s description of the fallacy is not analogous to what the Multiverse Objection says. A different analogy would suit us better.

Suppose, instead, that you and I walk into a casino and fall asleep in two very comfy beds. Suppose further that before we fall asleep we are told that we will only be awoken if somebody rolls ten sixes in a row on a fair die. When we wake up, I look at the time, note that we have only been asleep a couple of minutes and declare ‘Wow, the casino must be full tonight’. Unlike Goff’s example, the claim about the casino being full seems entirely reasonable. If there are a lot of people gambling in the casino, the likelihood that we will wake up so quickly becomes a lot higher. This is, I think (my anthropics are very rusty), an application of the Self-Indication Assumption - an anthropic principle for which you can find lots of arguments on ’s page.

This is more like what the Multiverse Objection is saying. In my casino, with the comfy beds, we don’t consider ‘What were the odds of our gambler rolling ten sixes?’ - any gambler rolling ten sixes will do, so it doesn’t matter who did it. We are awoken whenever anybody rolls ten sixes. Likewise with a multiverse, it doesn’t matter which universe randomly gets assigned the life-permitting constants - we just need one to get the constants, and then we will begin to exist in that universe, simply because it is the only one that would permit us to exist. Therefore, it seems clear to me that the Multiverse Objection does not commit the inverse gambler’s fallacy. It is in fact entirely reasonable to argue that the FTA gives evidence for the multiverse2.

However, even though the FTA gives us evidence for the multiverse, that does not mean it does not give evidence for theism. It is entirely possible for one datum to increase the likelihood of two distinct hypotheses. Swinburne discusses this in the opening chapters of ‘The Existence of God’ - he encourages us to imagine a crime scene. Suppose footprints are found which match the shoes of Suspect A. It would be silly for one detective to tell another: “The footprints aren’t evidence for Suspect A committing the crime, because they also fit the shoes of Suspect B!” No, it is not all or nothing, with the two hypotheses competing for the evidence. The evidence in fact increases the likelihood that Suspect A committed the crime, while simultaneously increasing the likelihood that Suspect B committed the crime. Sure, it would be even more likely that Suspect A commmitted the crime if we could rule Suspect B out, and similarly, the FTA would be even better evidence for theism if we could rule the Multiverse out, but they are both still good evidence for the former hypotheses in each case.

MFW I realise that the Eventually Response isn’t even necessary for the FTA to still be good evidence for theism.

To conclude, then, I think Epsilon’s NBLO against the FTA fails, and the Eventually Response, which he gives on the theist’s behalf, does not committ the inverse Gambler’s fallacy.

Epsilon wrote:

Bayesian arguments from evil usually hinge on this premise:

P(E|G&B) < P(E|¬G&B)

That is, the probability of evil given that God exists in conjunction with our background evidence prior to taking into account E, is lower than the probability of evil given that God does not exist, relative to the same background evidence. Put differently, evil is less probable on theism than on non-theism, relative to B.

A common response is to argue that we are completely in the dark and in fact have no idea how likely evil is given theism. Since our cognitive faculties are limited, we are ignorant of the reasons God could have to permit atrocious suffering in the world. That is to say, P(E|G&B) is inscrutable to us, or that P(E|G&B) ranges between 0 and 1 and we have no clue what it is. Then for any evidence X: P(X |G&B) should likewise be inscrutable. If we have no access to what God would bring about, then no evidence can raise or lower the probability of theism. Therefore, this response plausibly renders theism evidentially inert. For example let FT stand for fine tuning: P(FT |G&B) would be just as inscrutable to us. To sum up, this response to Bayesian arguments from evil could undermine a good chunk of arguments for God.

Now for my response:

I will argue that the two arguments (the fine-tuning argument and the inscrutability response to the problem of evil) are in fact consistent – while the design argument makes a claim about what God would aim at, the inscrutability argument calls claims about how God would achieve his goals unwarranted.

The putative inconsistency is clear thanks to Epsilon’s framing. Whereas the theist who argues for fine-tuning posits that God has an interest in human life, the sceptical theist seems to argue that God’s goals cannot be known. Thus, it seems that the theist can make one of these arguments, but not both.

However, further examination of the claims made in each argument will show that they are not in fact inconsistent.

The fine-tuning theist claims that God would be interested in the existence of human beings. Why might they suggest this? Firstly, since God is all-good by definition, God must pursue goodness. As Alvin Plantinga argues (in The Nature of Necessity, 1974), moral goodness is something God would pursue, and moral goodness consists in free beings choosing to do the right thing. Therefore, God should aim to create humans.

There are two versions of sceptical theism/inscrutability response, one of which conflicts with the design argument, one of which does not.

The first version of sceptical theism argues that God’s goals are inscrutable. Because we don’t know what God is aiming at, what the good looks like, we cannot say that God would not allow certain evils.

The second version of sceptical theism argues that God’s methods are inscrutable. While we can know that God would desire goodness in his universe, and we can have some ideas about what that might look like, the way he achieves this goodness is mysterious. This is the correct way of spelling out sceptical theism, as it does not conflict with the proposition above that God desires moral goodness.

However, should we support the second version of sceptical theism over the first? It might be argued that the first version of sceptical theism is more plausible. This is because goodness (God’s goal) is a concept of which we have no firm grasp (Murray, Nature Red in Tooth and Claw, 2008), hence the endless wrangling over the concept in moral philosophy. Additionally, on theism, goodness is plausibly grounded in the nature of God, according to many divine command theories, and since God’s nature is ineffable, the good should also be something of which we cannot get a reasonable grasp.

Nonetheless, the theist need not suppose goodness to be a mysterious concept. Firstly, God reveals goodness through divine revelation. The life of Jesus, and the commandments He gives provide a Christian theist with a reasonable account of what the good looks like, and other non-Christian theists have similar ways of accessing the good, e.g. the life of Muhammed. Secondly, the theist might suppose that God has ‘written his law on our hearts’, as Paul writes, giving us reliable intuitions about what good and evil look like. Lastly, it is worth noting that the argument I hinted at above, that God’s nature is ineffable and goodness is grounded in God’s nature, and thus goodness is ineffable, need not be considered valid - it is not at all clear that ineffability is a consequence of being grounded in an ineffable thing.

I have shown that the theist need not suppose goodness, and hence God’s goals, inscrutable. What reason is there to suppose that his methods are inscrutable? There are many reasons, and I have only space here to detail a few. Firstly, we don’t know about the necessary connections between good and evil in the universe. It is not easy for us to know what is metaphysically necessary and what is metaphysically possible (as arguments surrounding the ontological argument show). Secondly, we don’t know about the nature of the human soul, which, on many plausible theisms, is a complex part of human nature. As Hick (in Evil and the God of Love, 1966) suggests in his ‘soul-making theodicy’, there may be features of the soul that are refined and made more perfect by suffering, in ways we cannot understand. Lastly, we don’t know about the distant past and the future, including the afterlife and heaven. Goods and evils exist in the afterlife, and are plausibly connected to the goods and evils we see around us today, but the exact nature of these connections is mysterious. Thus, the connections between goods and evils is mysterious, and so God’s methods are mysterious. This makes the second version of sceptical theism plausible.

Thus, it seems there is no inconsistency between a fine-tuning argument and an inscrutability response to the problem of evil. While the fine-tuning argument makes a claim about God’s goals, a claim about what goodness consists in, the most plausible version of the inscrutability response argues that claims about God’s methods for achieving his goals are mysterious.

Epsilon might choose to respond by pointing out that the fine-tuning argument makes a claim about methods, which the version of sceptical theism I have proposed says is unknown to us. This would be that God uses the free actions of humans as His method to achieve moral goods.

I grant that the fine-tuning claim about God’s goal of creating humans could be reframed so as to seem more like a claim about His methods. However, I don’t think this is a problem for the sceptical theist. While the sceptical theist says that God might have some methods which are inscrutable to us, the sceptical theist does not need to say that all of God’s methods are inscrutable to us. In fact, it seems that this one method is very scrutable to us - how could God generate moral goodness without free actions? It seems to be a clear impossibility. Therefore, my response on the behalf of the sceptical theist stands.

In the next Everything Evil post, I will return to the line of argument started in part I and discuss how different versions of sceptical theism interact with different versions of the evidential argument.

Before shattering them, we must set up these binaries. Even though we will ultimately reject them, they will be instructive and helpful for understanding Lewis. The first political binary is the modern vs medieval divide often described by theologians writing under the banner of Radical Orthodoxy. I am using Simon Oliver’s introduction from ‘The Radical Orthodoxy Reader’ (2009) as my foundational text here. The second is a more theologically motivated divide, that of Eros vs Agape. When discussing this second divide, I have in mind the work of Roy Hattersley in his book ‘Socrates and Jesus: The Argument that Shaped Western Civilisation’ (2009). It is a distinction in theology of love, but Hattersley argues that it affects political philosophy dramatically. As we shall see, the two categorisations are, in my eyes, coextensive, but since their originators seem to disagree with me on the details of this1, I will initially keep the two separate.

Medieval vs Modern Political Philosophy

What is the medieval-modern divide? For our purposes the most important difference is in how we understand the infamous state of nature. For medieval Christian political philosophy, according to Oliver, society is a natural phenomenon. You may have learned about The Great Chain of Being in GCSE History when discussing the Elizabethans and their ingrained ideas of hierarchy. This is exactly the mindset Oliver describes as medieval: man is subject to nobles, who are subject to the King, who is in turn subject to God. Violence, sin, and conflict in general are seen by the medieval thinkers as an inruption into this naturally peaceful order.

For the moderns, on the other hand, most notably exemplified by Hobbes and his Leviathan, the state of nature is one of disorder, self-interest, and violence. Thus, the modern political philosophy takes the framework of medievalism and flips it on its head. Society becomes an unnatural force imposed on nature in order to correct its flaws. Moderns such as Rousseau and Locke see society as a product of human cooperation to reach some common goal. First come individuals, and second comes the society they use to govern individuals. For the medievalist, this was completely wrong - for them, first comes society, and second comes the individual who finds his self-identity in his social role.

Eros vs Agape - theology of love shapes politics

Hattersley’s work paints Eros and Agape as two entirely different worldviews, rather than mere political philosophies. Together they are, he argues, the engine of Western civilisation. I cannot detail all the differences between the two here (without keeping you forever), but the distinction important for our purposes concerns the relationship between God and man. For followers of Eros, initially the ancient Greeks, but also many later Western thinkers, the relationship between God and man is a continuum. Man starts in ignorance and mere humanity, but through his appreciation of beauty - hence the term Eros, relating to erotic love - and through his imitation of that beauty, man can come to imitate God more greatly.

So much for Eros. What of Agape? Followers of Agape are mainly Christian theologians, with Luther perhaps being its fiercest proponent. For these thinkers, there is no continuum connecting God and man. Instead, there is a radical divide. Man is incapable of traversing this divide on his own. Thus, God must reach down to man and gift him a greater participation in the divine. Hence the Incarnation - Jesus becoming man and instituting the sacrament of baptism allows man to be united with the Father.

What are the political consequences of these contrasting theologies?2 Eros, in its affirmation of man’s potential self-subsistent divinity, paves the way for democracy; in a democratic system, the value of each man is recognised by his share in political power. Man reaches upwards to create government. Contrasting this, Agape’s conception of man as fallen requires something greater than him to reach down and uplift him. In medieval times this was the divinely ordained monarchy, and in Soviet Russia it was the all-powerful state.

You are perhaps already beginning to see how the two categorisations of politics I have described map onto one another. The modernist version of political philosophy, wherein the state is the creation of individuals, pairs well with Eros theology which validates democracy. Here, man is seen reaching upwards towards divinity. The medieval version pairs with the more hierarchical politics of Agape. In this system, the state, which is natural and divinely ordained, helps fallen man to be greater3.

Subscribe now

Thus, we have a two-part typology of political philosophy which is entwined with theology. We might think that C. S. Lewis, ever awake to the theological implications of his views, would take note of this division and fall clearly into one camp. And if you selectively read Lewis’ work, the case for this is very easy to make:

Lewis as a medievalist

In Lewis’ essay ‘Membership’ (1945, included in ‘The Weight of Glory and other addresses’, 1976), Lewis immediately sets himself apart from modernist accounts of society by distinguishing membership in a set from membership in a body. Membership in a set is, for Lewis, the more secular modern view of society, wherein each member has their individuality before and independently of their role in society, and the society serves the purpose of propping up those individualist lives. Membership in a body, however, is the more involved and, Lewis implies, superior way of understanding a society4. In this way of viewing society individuals are more than just part of a set - they are more like organs in a body, each with a unique and separate function or role. Moreover, members are not individuals with personality separate from the body. Instead, their personality is discovered as an outworking of their role in the body - they are made for their purpose rather than their purpose being made for them. Thus, it seems Lewis wholly dismisses the inorganic nature of modern-Eros politics, such as Hobbes’ Leviathan, where people construct society. Instead, for Lewis, society constructs people.

Other works by Lewis paint a similar picture. In ‘The Four Loves’ (1960), we see that Lewis considers Agape, which he calls Charity, the highest love, and even though Eros can set itself up as a god and is followed by many men, all loves including Eros must be tempered by Agape. Lewis recognises, as the ancients did, that loving somebody through Eros can help us to become more and more like God, loving another as ourselves, fulfilling the law with regard to one person. However, this resemblance to God-like love is not complete and will not be extended to all people without the influence of Agape. Thus, Lewis clearly prioritises Agape, which further supports the view that he is a medievalist when it comes to politics.

The last piece of our portrait of Lewis as a medievalist is the society he describes in ‘Out of the Silent Planet’ (1938)5. Spoilers ahead! As Malacandra, or Mars, is without sin, its inhabitants, the hnau, thus live in the way intended by God (or Maleldil), so we can see what Lewis’ idea of a perfect society is like. While Hobbes and other modernists have a coercive state which prevents men from pursuing their own self-interest at the expense of others, Malacandra has no state. Instead the hnau obey the authority of their Oyarsa out of recognition of his divine authority. Hnau society is extremely hierarchical and organic, with the different species on Malacandra fulfilling distinct functions for which they seem designed. Lewis is showing his vision of hierarchy without domination, where all in the hierarchy live harmoniously and pursue the good of others. Thus, again, it seems that Lewis is supporting the medieval-Agape conception of society. However, thus far in my analysis of Lewis I have only been telling one side of the story. There is more to Lewis’ political philosophy than I have thus far let on.

Lewis as a democrat

Despite all the evidence for Lewis’ medievalism, a fuller reading of his essays ‘Membership’ (1945) and ‘Equality’ (1943) reveals something more interesting. In the latter, Lewis straightforwardly declares “I am a democrat because I believe in the Fall of Man”. Why this seeming contradiction?

The answer is in Lewis’ understanding of nature. Even though Agape theology understands the natural state of affairs to be one of a peaceful hierarchical society which is disrupted by sin, Lewis adopts a more Augustinian understanding of nature wherein sin has so occupied man’s soul that he has a different functional nature - one with a will to power and domination. As such, for fallen men, in Lewis’ eyes, the natural state of power is a corrupt one.

Agape might be the supreme form of love, and one which would order the perfect society, but since men are so out of touch with God and His perfect Agape-love, they are incapable of bringing this society about. Lewis writes in ‘The Four Loves’ that he is not even sure he has been able to ever truly experience Agape-love from himself towards another. As such, society must be ordered by different rules than mere Agape.

This is where Lewis’ appreciation for equality comes in. Even though he considers it a fiction, as men are by nature unequal, Lewis thinks it a useful fiction, instrumentally good as a medicine for our sickness of domination and corruption. Political equality, in the form of democracy, and economic equality, which Lewis says we need more of (in ‘Equality’), are good defences against evil in those spheres of life. They are like clothes - necessary since the Fall in our social lives, but not in our personal and spiritual lives. Lewis is in favour of hierarchy in the home and in the Church, just so long as this does not develop into political or coercive control that can be abused.

This conception of equality as a bulwark to stop evil men rather than a natural state is seen clearly towards the end of ‘Democratic Education’ (1944) where Lewis describes how less intelligent men can take pleasure in disobeying the rules of would-be tyrants who are more intelligent than them, thanks to their political equality.

It seems, then, that Lewis has a conception of Agape influencing his politics which is even more radically Christian than the medieval view. Rather than the state pulling man up from his fallen position because the state is an image of God, as the medieval view held, Lewis considers the state and the men in power to be fallen too, meaning they cannot help raise other men up. I say this is more radically Christian because it emphasises even more how man is helplessly corrupted without God. Furthermore, instead of men working together to become more divine and good, as Eros-modernism suggests, Lewis sees the main goal of society as preventing other men from sinking down to more and more evil ways of life. Instead of man becoming more divine, democracy for Lewis is men preventing other men from becoming more demonic. It is up to God and God alone to help us to become divine. This understanding can be seen clearly in the sequel to ‘Out of the Silent Planet’, ‘Perelandra’ (1943). Spoilers ahead! In ‘Perelandra’ it is not Ransom’s job to help the Adam and Eve of Venus to reach God, but instead it is his role to prevent them from falling into sin. Ransom fights Weston to extinguish evil from the fledgeling society on Venus, and it is only then that Maleldil initiates the uptake of the Adam and Eve of Venus into a heavenly society.

Lewis’ political philosophy also neatly fits in with his eschatology. Instead of men working to bring about God’s kingdom, as postmillenialism would have us imagine, Lewis thinks that God’s judgement will come at an unforeseeable time and interrupt history, rather than take place as an ultimate natural progression of history. This is also seen in ‘The Last Battle’ (1956), where society gets significantly worse, with a false corrupt ruler before Aslan’s final rule over Narnia. Thus, even though Lewis is in favour of political equality, he definitely does not share the same vision of political progress that modernists imagine might arise from it. Progress in this sense is also rejected by Lewis in other works such as ‘The Abolition of Man’ (1943) and ‘Willing Slaves of the Welfare State’ (1958).

Conclusion

To conclude, even though the binaries set up by Oliver and Hattersley are very informative for analysing political philosophy, Lewis does not fit snugly into either camp. Lewis is more focused on how man can prevent evil in politics than he is on how man can do good, and so he is in favour of political equality. Nonetheless, he thinks that the natural (pre-Fall) state for man is one of hierarchy. As such, hierarchy is part of our psyche and helpful in personal and spiritual fields, but it is too corrupting to be used in the political arena. While I researched for this post, Chat-GPT made up many an apocryphal letter and quote by Lewis, and one, although hallucinated, seemed a particularly accurate understanding of Lewis’ view: “Democracy for fallen men; monarchy for heaven.”

The Logical Problem of Evil

I'll start with the most ambitious argument. The logical problem of evil argues that the existence of evil is inconsistent with the existence of God. Therefore, if evil exists, as we seem forced to admit, God cannot exist as well. This argument, then, threatens to definitively prove that God doesn't exist.

The inconsistency touted by proponents of the logical problem isn’t strictly a logical inconsistency between evil and God’s existence. As Plantinga notes2, there is obviously no principle of formal logic which makes the following two propositions incompatible:

(1) God, an omniscient, omnipotent, and omnibenevolent being, exists

(2) Evil exists

However, if the proponent of the argument can put forward another proposition, such as

(3) If an omniscient, omnipotent, omnibenevolent being exists, then evil does not exist,

then the trio of propositions are logically incompatible. Therefore, if we think (3) is true, and the proponent of the argument will argue it is (perhaps it is even an analytic truth), then (1) and (2) become inconsistent with one another. So, we have to either deny that God exists, or deny that evil exists; since evil is, to many, self evident, it seems the only reasonable option is to drop (1) and deny God’s existence.

Why should we think (3) is true? The thought is quite simple - if God were omniscient, He would know about all evil that could occur; if He were omnibenevolent, He would want to eliminate all of that evil; and if He were omnipotent, then He would be able to eliminate all the evil. Thus, if He existed, he would eliminate all evil.

Note that because the logical problem is so ambitious, all the theist needs to do to rebut it is provide a possible way in which God and evil might be reconciled; they need only show that the entailment in (3) is not a necessary entailment.

Theodicies

Usually at this point come many theodicies which attempt to show why (3) is false. Theodicies look at evils that actually occur and attempt to explain the goods that come from them in order to justify God in allowing those evils to occur. This is in contrast to a defence, which doesn’t claim to have a justification for the actual evils we see, just to provide a possible way in which God could exist alongside evil - to show that (3) is not a necessary entailment. Thus, strictly speaking, a theodicy is not necessary to defend against the logical problem of Evil. Only a defence is necessary. Thus, I will not linger on theodicies long, but will just point out two general problems.

First, theodicies usually have limited scope of explanation. For example, the secondary goods theodicy only applies to evils which humans3 can respond to with a virtue, leaving out evils which humans cannot sufficiently respond to like the death of a fawn in a forest, not known of by any human. Thus, unless armed with a vast array of theodicies to cover every evil known to man, the theist who attempts to respond to the logical problem will be better served by a defence rather than a theodicy.

Secondly, any theodicy has to face off against the ‘more omnipotence’ response from the atheist. The point here is that any theodicy will have to point out a limit to God’s omnipotence, saying ‘God can’t actualise a particular good without allowing an evil’. A veteran online atheist will see ‘God can’t’ and immediately jump at the bit to point out that this is a limit on God’s power, and so the theodicy fails to save our traditional view of an omnipotent God. The ‘more omnipotence’ response, then, is to respond to a theodicy by saying “But surely a God with more omnipotence than that would be able to make the good without the evil! And a God with less omnipotence is no God at all!”

Contrary to the veteran atheist, however, it is clear that there are some real limits to God’s omnipotence, which don’t necessarily prevent us from rightly calling Him omnipotent. For example, it seems reasonable to many philosophers, to say that God is bound by the laws of logic, so He cannot do things like create a square circle. Omnipotence would then mean the ability to do anything which is logically possible. Further, it might be right to say that God cannot create things which are metaphysically impossible4, such as a ball that is green all over and red all over. Omnipotence would then just mean the ability to do anything which is metaphysically possible. This seems a respectable alternative definition of omnipotence. Therefore, it could be reasonable to suppose that there are certain goods God cannot actualise without actualising certain evils, because there is a logically or metaphysically necessary connection between the two. Therefore, theodicies, while they must always face off against the ‘more omnipotence’ response, do not necessarily flounder as a result.

The Free Will Defence

As noted, though, a defence is more apt to solving the logical problem than a theodicy is.

The Free Will Defence, attributable to Plantinga5, says that (3) is not a necessary entailment, because it is possible that God cannot eliminate all evil, or at least it is possible that He cannot do so without also eliminating the possibility of good. This is because in order for moral goods to be possible, moral agents must be free to decide whether to perform good or evil - if the decision to do good isn’t made freely, then it isn’t a moral good. Therefore, for there to be moral good in the world, God has to create free agents. If agents are free, it is not up to God whether they decide in one moment to perform the right action or the wrong action. Thus, if God grants freedom, He can’t use His omnipotence to actualise just any particular world - instead He allows agents to actualise certain worlds using their freedom. That means God can’t use His omnipotence to prevent all evil, without robbing the world of moral good, which an omnipotent God would presumably not do.

The atheist response here may be to point out that even though actions are free, there is intuitively still a fact of the matter about whether somebody would perform a good or an evil when given the free choice. Thus, God can know, for example, that the BBC execs would, in the circumstances which actually occurred, cancel Doctor Who in 1989, a great moral evil. He could then decide not to actualise the circumstances in which this occurred, and instead actualise ones in which He knows that the execs would not cancel Doctor Who. Therefore, God can allow agents the freedom to perform moral goods, while being sure that they would not perform evils.

Plantinga’s retort here is to suggest that people suffer from something he calls ‘Transworld Depravity’. If someone is depraved transworldly, they are such that no matter which world God actualises with them included, they perform at least one moral evil. Since it is at least possible that all possible people are depraved transworldly, Plantinga has succeeded in providing a way in which the co-existence of God and evil may be possible6.

Problems for the Free Will Defence

Yet there is a problem for the Free Will Defender, one which Plantinga does not deal with. This is the ‘more omnipotence’ response, as we have already seen with theodicies: if all the people that could be created suffer from transworld depravity, why is it that God can’t just create different people, or edit the people that already exist, such that they don’t exhibit transworld depravity? Surely an omnipotent God, one who is meant to be the creator of people, can do this! To argue that God cannot do this seems to suggest a picture of creation where God has vials filled with uninstantiated souls, which already existed without his handiwork, and which he can choose to instantiate in real people or leave in the vial. These souls either do or do not have transworld depravity and God cannot change that fact.

Nonetheless, as noted before, it seems reasonable to suppose that there are some things which are logically or metaphysically necessary, and so God cannot change them. The fact of transworld depravity may be one of these things, and so maybe we can say that God’s omnipotence is not unduly limited by the Free Will Defence. But it seems unclear to me why the transworld depravity of a person would be logically or metaphysically necessary. Thus, we need some more argumentation for this to be a wholly satisfactory response to the logical problem.

Another difficult question for the Free Will Defender concerns the nature of free will. So far it has been supposed that free will entails the metaphysical possibility of evil actions as well as good. This is libertarian free will. But what about compatibilist free will? A compatibilist says somebody has free will if their actions are a result of their desires and intentions. This compatibilist version of free will doesn’t entail that the decision made is metaphysically contingent - instead the decision might be metaphysically necessary, wholly determined by physics, but still free, according to the compatibilist, because a result of desires and intentions. It is not clear, as J. L. Mackie7 argues, that incompatibilist freedom is the kind required for us to perform moral goods. Instead, compatibilist freedom might be enough. If this were allowed, the result would that God could in fact create beings who performed moral goods without the metaphysical possibility of them also performing moral evils. Therefore, the Free Will Defence would fail.

How could the Free Will Defender respond? I do not have time here to consider a full response, but one might include an explanation of why compatibilist freedom is actually no freedom at all, or why moral goodness requires incompatibilist freedom in particular.

To conclude this section, the logical problem of evil is not, contrary to popular narratives, dead in the water due to theodicies and the Free Will Defence. Both methods of responding to the logical problem face particular issues, which are not clearly resolved in their favour.

Evidential Problems of Evil

Nonetheless, many philosophers consider the logical problem to be the less promising version of the argument, and instead put forward the evidential problem. The evidential problem does not posit that the existence of evil is inconsistent with the existence of God, but simply that the former provides evidence against (or disconfirms) the latter. Thus, if this argument succeeds, we have strong, perhaps even conclusive, evidence against the existence of God.

Plantinga, hot off the heels of offering the Free Will Defence against the logical problem, tries his hand at responding to the evidential problem, or, as he calls it, ‘The Probabilistic Argument from Evil’8. Plantinga’s response is too wordy to quote here in a way that makes sense to the reader, so I will paraphrase.

Consider the propositions, as above:

(1) God, an omniscient, omnipotent, and omnibenevolent being, exists

and

(2) Evil exists

We should further consider, Plantinga argues, the following conjunction:

(4) Evil exists and God, an omniscient, omnipotent, and omnibenevolent being, exists.

Now, (2) does not disconfirm (4). Plantinga continues, if a proposition, p, confirms or disconfirms another proposition, q, and q entails r, then p also confirms or disconfirms r. (4) entails (1), and so (2) does not disconfirm (1), because it doesn’t disconfirm (4).

Hopefully the mistake here is clear despite all the numbers and letters. Plantinga is right that if p disconfirms q and q entails r then p disconfirms r. However, this is not the inference rule which Plantinga needs for his conclusion that (2) does not disconfirm (1), that evil does not provide evidence against the existence of God. Instead he needs it to be the case that

(IR) If p does not disconfirm q, and q entails r, then p does not disconfirm r.

However, (IR) is simply false. Against it, consider the following propositions.

(5) He voted Conservative

(6) He voted Conservative and he is a member of the Labour party

(7) He is a member of the Labour party

Now, (5) obviously does not disconfirm (6), because (6) includes (5). Further (6) entails (7). But, it does not follow therefrom that (5) does not disconfirm (7). In fact it does - if somebody voted for the Conservatives, it becomes unlikely that they are also a member of the Labour party. You can easily come up with dozens of cases like this - just take a proposition that disconfirms another and jam the two of them together in a conjunction. Every time, you’ll end up with a counterexample to Plantinga’s principle, (IR). Thus (IR) is not true, and so Plantinga’s response to the evidential problem fails. He has not shown that (2) does not disconfirm (1).

Subtypes of the Evidential Problem
Thus far the discussion has dealt with the evidential problem in general. However, there are several different types of the evidential problem. Each evidential problem holds that we have some evidence for the existence of gratuitous evil - evil that God would not allow, were He to exist. They differ in how they reach the point where they can reasonably say they have evidence for this gratuity premise. I thus divide the evidential problem into three subtypes: the Noseeum problem, the Calculation problem, and the Commonsense problem. This categorisation will be useful in future blogposts where I discuss different methods of resolving the evidential problem - as we will see, some solutions will work against some subtypes, but not others.

The Noseeum Problem

This version of the problem takes premise 8 as granted:

(8) There is inscrutable evil,

where inscrutable evil is evil for which we can’t see/understand reasons that would justify its existence in a world created by a tri-omni God. From premise 8, a Noseeum inference is made: “Regarding reasons for this evil, I don’t see ‘um, so they ain’t there!” This leads to the conclusion that there is evil which doesn’t have any justifying reasons, i.e. evil which is gratuitous.

As we will see in future blogposts in this series, the use of a noseeum inference is questionable. Sometimes noseeum inferences are valid. For example, when considering if there are elephants in my garage, the fact that I don’t see any is a good reason to suppose that there aren’t any. Sometimes noseeum inferences are invalid - when considering if there are insects in my garage, the fact that I don’t see any when standing outside looking in is not a good reason to suppose that there aren’t any9. The question that we will face in future posts is: are God-justifying reasons for evil more like elephants in the garage or insects in the garage?

The Calculation Problem

This is perhaps the most implausible version of the problem of evil, one which I have not seen defended by any philosophers. This version of the problem starts with inscrutable evil (premise 8, as above) and argues that, even though the evils are inscrutable in the sense that we can’t see any justifying reasons for them, they are not inscrutable in the sense that we can’t figure out whether there would be justifying reasons. The calculation problem, then, involves calculating the consequences of the evil, weighing up the goods and bads, and concluding that the evil is (probably) gratuitous.

For this method to be plausible, we would need great abilities of insight into necessary connections between goods and evils and into the way goods and evils should be weighed. Most would consider the idea that we have these abilities preposterous (although that does not stop some of our good utilitarian friends from acting as though they have such powers). Therefore, the calculation problem will be the problem most easily done away with by theists.

The Commonsense Problem

This version of the problem does away with the preliminary step of premise 8, and instead jumps straight to the gratuity premise (There is gratuitous evil). According to proponents of this problem, we have commonsense faculties which allow us to see, without need for arguments or evidence10, that certain evils are gratuitous ones.

Although the ability to identify gratuity in this way is contestable, like the calculation abilities touted by the proponent of the calculation problem, commonsense methods of inference are more respected philosophically, especially by externalists and commonsensists about knowledge. As we will see in a later blogpost concerning the problem of evil and general scepticism, these schools of thought are ones which theists are likely to avail themselves of in attempts to avoid epistemological problems. But adopting either one of these schools of thought, and thereby accepting many commonsense inferences as justified, requires the theist to then explain why the commonsense inference to the gratuity premise, made here, is not justified. Thus, the commonsense problem is a greater threat to the theist than the calculation problem, and arguably than the noseeum problem, too.

Enough, now about the evidential problem. There will be a lot more to come in future posts!

The Modal Problem of Evil

The modal problem originates with Theodore Guleserian in 198311. Guleserian points out that a necessarily existent tri-omni God is incompatible with some standard modal views. Namely, it is standard to suppose that there are possible worlds containing great evils, such that it would be impermissible for God to actualise those worlds. But if God is omnibenevolent (or morally perfect), He would not actualise any world which it was impermissible for Him to actualise. Therefore, those worlds are not actually possible if a tri-omni God necessarily exists. To maintain our standard modal view, then, that those worlds are possible, we must sacrifice one of the four properties we typically assign to God - those being the three omnis (taken as essential properties of God) and His necessary existence.

Now for my original response to the modal problem. I think the theist should bite the bullet and drop one of God’s necessary omni-properties, namely His necessary omnibenevolence/moral perfection. Typically, philosophers of religion argue that the necessary possession of a great-making property such as moral perfection is greater than the contingent possession of such a property. I think, however, that this intuition is wrong in the case of great-making moral properties. If one performed good actions necessarily, it would seem that one didn’t really have free will regarding these actions, at least not in the libertarian sense laid out above. As such, it seems that if one performs a good action contingently, then one is morally greater than if one had performed it necessarily. Thus, to describe the greatest possible being, God, we should hold that He is only contingently morally perfect. Thus, there are possible worlds which He possibly actualises, but commits a moral evil by so doing. It is only in this world, the actual one, that God is morally perfect, as He actualised the best world He possibly could.

You might see this response and think that it is not truly a response, but an acceptance of the conclusion of the modal problem, that there is no necessarily tri-omni God. And you would be right. But, I hope that in the way I have accepted the conclusion I have elucidated a way that this acceptance might be seen as a refinement of the theistic position rather than as a reason to reject it altogether.

Conclusion

I have described several versions of the problem of evil, and hinted at the ways in which they are better and worse than one another. The logical problem, though considered dead by many, is not categorically defeated by the Free Will Defence - the Free Will Defender still has to answer questions about ‘more omnipotence’ and the nature of free will. The evidential problem is not defeated by Plantinga’s response. Its subtypes, the noseeum problem, the calculation problem, and the commonsense problem, interact with responses such as skeptical theism in different ways, which we will see in upcoming blogposts. Finally, the modal problem shows us how our understanding of God needs to be refined in order to avoid contradiction.