A summary and discussion of The Problem of Pain by C. S. Lewis, ISBN 9780060652968. (The text is available online in various formats, for instance here.) The book is about the problem of the existence of human suffering in the presence of an omnipotent and good God, often known as the problem of evil.

I picked up this book because of my long-standing interest in understanding why relatively many contemporary intelligent people are religious. I’m an atheist (although, I hope, not an annoying Internet atheist); my impression is that religious people¹I specifically mean people who posit the existence of a God who has an influence on the material world, and usually one that can be influenced by humans or is specially directed towards them. I have no particular issue with deists. I also don’t mean people who participate in organized religion for social or personal growth reasons, maybe because they think religious texts or priests are particularly wise, but are not believers. are committing an egregious epistemic error, which they then cover up with fantastical explanations when the cracks (which are actually not always epistemic in nature) show. The problem of evil seems to be one of these cracks, so I was curious to read about an explanation by a Christian. Plus C. S. Lewis is a former atheist, which makes the book all the more relevant to me.

Summary

This section is my best attempt at rendering of the arguments in the book. I don’t necessarily endorse the ideas. I think some are actually pretty awful if you apply them to real cases.

C. S. Lewis summarizes the problem as:

If God were good, He would make His creatures perfectly happy, and if He were almighty He would be able to do what he wished. But the creatures are not happy. Therefore God lacks either goodness, or power, or both.

Being a Christian book, the author doesn’t try to make that case that God being good and omnipotent is the most parsimonious explanation. Instead, he assumes these qualities of God, and then explains how they are consistent with the existence of suffering.

First, he remarks that God has given humanity free will. The author notes that this in itself implies the possibility of pain in the world: if people can act as they choose, they can choose to hurt others. Creating a world with free will but without pain is a logical inconsistency, so an impossibility even for God. (C. S. Lewis rightly notes that this doesn’t mean God is not omnipotent: omnipotence doesn’t cover logically impossible things). However, he declines to judge or justify the fact that God has chosen to create a universe with free will. We must assume God had good reason, even if it is not apparent or comprehensible to us.

But if people can do whatever they want, they must choose to renunciate their self-centering and submit their will to God’s:

Christian renunciation does not mean stoic “apathy”, but a readiness to prefer God to inferior ends which are in themselves lawful. Hence the Perfect man brought to Gethsemane a will, and a strong will, to escape suffering and death if such escape were compatible with the Father’s will, combined with a perfect readiness for obedience if it were not. Some of the saints recommend a “total renunciation” at the very threshold of our discipleship; but I think this can mean only a total readiness for every particular renunciation that may be demanded, for it would not be possible to live from moment to moment willing nothing but submission to God as such. What would be the material for the submission? It would seem self-contradictory to say “What I will is to subject what I will to God’s will”, for the second what has no content.

I admit that after reading that, I’m still a little fuzzy on what exactly “submitting your will to God” means. If that helps, here’s what not doing it is:

The movement whereby a creature (that is, an essentially dependent being whose principle of existence lies not in itself but in another) tries to set up on its own, to exist for itself.

Of course, this is what we are obligated to do with free will, not the only thing we can choose to do (we can choose to do whatever we want, or else free will wouldn’t be free). This obligation is not just an arbitrary order from God: God commands things if and only if they are good. So submitting their will to God’s is the only way humans can do what’s good, be truly happy, and fulfill their purpose.

But submitting to God, as an act of self-renunciation, is always a little painful. So of course, humans aren’t doing what they should be doing, and are never submitting to God. The events of genesis describing the original sin are really an allegory for primordial humans converting from the God-centered free will to the self-centered one. In particular, this fall is not something we need to expect to be universal among all species that have free will.

Admitting there are these two different modes of using free will (and since we’re always using free will, effectively two modes of existence), the thesis of the book is that the existence of pain is one of God’s ways to help us move from the natural-since-the-original-sin self-centered mode of operation to the salvationary God-centered one.

According to this idea, it’s easy for people to get stuck in a sort of local optimum where they feel content or happy with their lives, and stop thinking about God. But as we’ve seen, this self-centered use of the will can’t bring about real happiness, so God pushes you up the pain gradient so you’ll be incentivized to turn toward him and ultimately reach an even better condition.

More specifically, pain helps us travel on a path towards God:

• First, it makes us realize that, even if we were content, not everything is well; indeed, humans are naturally evil because they have fallen.
• Second, it helps us solve this problem. In some sense, pain is retribution for not having turned to God; in another sense, it is what helps us overcome our fault. Pain will force us to turn to God.
• Third, pain shows us that we are indeed doing things for God. If we only do things that we like, we can never know if we are truly acting for God’s sake. Hence in Christian theology, virtue ethics and deontology are not contradictory: what is good (deontology) is submission to God, and submitting to God is inherently painful (virtue ethics). For completeness’ sake, C. S. Lewis also mentions utilitarianism, to point out that it makes no sense. He argues that if someone suffers some pain level x, another person also suffering x does not increase the amount of pain, because nobody suffers 2x.

So pain is useful, and in fact necessary to the redeeming of humans. While we should still strive to eliminate pain as individuals, we cannot expect to ever eradicate it from the world, because that would be contrary to God’s plan.

Of course, pain sometimes fails to make people turn towards God, and instead has various consequences that seem worse than the cause. This is, once again, necessary in the presence of free will. If pain was guaranteed to cause submission to God, free will wouldn’t really be free.

In conclusion, I think a good high level summary of the book is that earthly life is meant as a way for people to find God, not as hedonist’s paradise, so the presence of pain isn’t surprising. Indeed, one of the final sections, which concerns Heaven, reassures us that our pains on Earth will come to a rapturous end there.

Hell

One of the ten chapters is dedicated to the problem of Hell. Hell initially seems a bit like adding insult to injury: not only do we have pain in the real world, but we also have a God-created place of eternal suffering for some people? (Of course, the blow is blunted by the fact that only bad people go to Hell).

The Bible seems rather light on details about Hell. C. S. Lewis suggests the proper interpretation is perhaps metaphorical, where Hell is the condition of no longer being a soul, or that of being a soul without God. (Souls in Heaven, in contrast, get to see God and submit to him all the time.) Okay, I guess.

Matt’s snarky comments

Before I delve into the complaints, I want to acknowledge some particularly good aspects of the book. First and perhaps most importantly, it was a pleasant and well-organized read.

Second, the claims made are generally compatible with scientific knowledge. The author makes no particular empirical predictions, and is an old-Earth evolutionist.²He’s possibly an intelligent designer, but that also doesn’t contradict empirical evidence, it’s just non-scientific.

Third, while the arguments are universal, many of the examples are a nice way to peer into early-20th century English society.

Now on to my problems with the book.

Free will

It seems like free will is somewhat frequently used in theodicies (i.e. solutions to the problem of pain), because it implies the possibility of pain between agents. If free will is extremely desirable, then maybe pain is just a natural consequence that doesn’t outweigh the benefits. Of course, this leaves the problem of natural pain (caused by the environment: disease, disaster, etc.) open, so in reality it doesn’t solve too much. While the book does present this argument, I think it’s actually not very reliant on it, which I appreciate.

However, free will remains a critical part of the explanation, because without free will, there can be no fall or original sin, and without original sin, there is no need for pain to set people right.

I have two main issues with this.

1. Why is free will so good? C. S. Lewis is insistent that divine morality is different from humans’, but only in the sense that it is more perfect: a reasonable person, faced with the God’s morality, would have no issue recognizing it is superior to their own. But the author refuses any hint at an explanation why free will is so morally desirable for God.

I would have liked at least an attempt, because if we accept the justification of “God is good, so we must assume there is a reason for free will to be necessary”, why can’t we also accept that there must be a reason for pain that we can’t know or comprehend? The book would be over in a quick 20 pages.

There are certainly various arguments you can make for the existence of free will. For instance, if God himself has it, having free will makes us more like God, which certainly seems good. But as it is, a key step in the argument showing a lack of contradiction between omnipotence, goodness, and the presence of pain, must be taken on faith.

2. I think the details of the posited model of free will might make a difference. If it’s incompatibilist–nondeterministic free will, then… nothing, nondeterministic free will makes no sense.

If it’s not, then God, who is omniscient, already knows who will finally submit to him and who won’t, who will make it to Heaven and who won’t, and knows the original sin would happen before he creates humans. Especially if we accept C. S. Lewis’ supposition that other sentient species might not have fallen, and thus free will was possible without pain (just create these unfalling species instead of humans), then this makes creating humans seem cruel even if we assume there was an excellent reason to desire free will.

Or maybe the solution is more unsettling. (This paragraph is very LessWrong-ian, so bear with me. Return to normality is expected in the next section.) Since we’ve talked about how logic bears on God’s omnipotence, maybe computing the outcome of a given universe is actually an irreducibly complex task, which requires a full simulation of said universe. Perhaps God can do this very quickly, or instantly, but doesn’t the mere act of “running the simulation” make the people in it exist, and suffer? Does God, by virtue of being omniscient, actually simulate (or has simulated) all universes, including the worst possible world?³If you like stuff with God creating multiverses, you might like Answer to Job by Scott Alexander, which I’ve discussed before.

Other conundrums

1. This seems unbelievably mean. According to this line of argument, terminal cancer patients, Holocaust victims, and abused children are to blame for their ills: they were simply too comfortable in life to get close to God without the spur of pain. C. S. Lewis acknowledges this, and says he’s only the bearer of bad news: this is the only way to make sense of the puzzle, so this solution must be correct.

But I think this is a strong argument against this whole theory. If I made a different theodicy that happened to prove that bananas weighted 18 tons, certainly that would discredit it. Doesn’t the conclusion that all terminal cancer patients are to blame, or at the very least are always in extremely dire spiritual straits, also discredit the argument at hand?

2. Why is pain the only way to get closer to submission to God? Couldn’t God, for instance, beam down very convincing information about him directly into our brains in a way that might even be pleasant? I didn’t necessarily expect a very long discussion of this, but I would have liked something.

3. When C. S. Lewis talks about the fall of man, he describes it as an individual experience, someone deciding to focus entirely on themself instead of God. So why does the fall apply to or affect all of humanity? Perhaps the answer is that this fall was caused by, or further caused, changes in social structures which then perpetuated this movement towards the self.

This seems related to Julian Jaynes’ bicameral mind theory,⁴Scott Alexander has a good review of the book. which posits that humans initially thought their inner monologue was the voice of gods talking to them, a situation that lasted until the bronze age collapse. At that point, trade (and the accompanying deception), as well as widespread societal collapse, forced people out of this mode of existence, and so entire societies’ mindsets changed radically.

The bicameral mind theory is not especially well supported, but maybe it provides a parallel to the way C. S. Lewis envisions the fall to work at a species-wide level.

4. This one is a total nitpick, but I don’t think the “nobody experiences the utility function” argument is very convincing. I know this is an argument about the foundations of morals, which I don’t think there is a great way to sort out. It’s still deeply weird to me to claim that it’s no worse for one person to suffer than for a million.

5. Also the book is sort of sexist. And sort of racist. But it’s really not the main focus, so I’ll have to blame that on the times.

6. This is an actual quote from the book (emphasis mine).

We forget that our prehistoric ancestors made all the most useful discoveries, except that of chloroform, which have ever been made. To them we owe language, the family, clothing, the use of fire, the domestication of animals, the wheel, the ship, poetry and agriculture.

I… don’t know what to say. This sentence would have been kind of fine without the bit about chloroform, but don’t you realize your argument is really temporary if you’re mentioning chloroform and live in a time when people are inventing electricity and airplanes? Or antibiotics? This seems like a really odd statement for a book from the 1940s.

Other assorted issues with this type of theodicy are presented in this section of the Stanford Encyclopedia of Philosophy.

Epicycles

Some ancient Greeks thought the Sun went around the Earth in a circle, a theory that was simple, had predictive power, and was very satisfying.⁵https://xkcd.com/365/

It was a bit more complicated for the planets, because they go around the Sun, so viewed from the Earth, they go around circles (corresponding to the Earth’s actual motion around the sun) whose centers go around circles (corresponding to the planet’s own motion around the Sun). This double-circle contraption is an epicycle.

But later on, people realized the observations didn’t quite match the model. In reality, planets go around the Sun in ellipses,⁶More or less, but bear with me. and in the late 17th century, the measurements to be modeled looked like this:

But that didn’t stop people! The now tried and true technique of adding epicycles could be applied one level deeper, adding one more circle on top of the circle that goes around the circle, to explain deviations from the simpler model. And you could go on adding epicycles to model deviations from the ellipse, perhaps due to relativity or measurement error.

The later development of Fourier analysis proved that any periodic trajectory where the planets don’t teleport, no matter how convoluted, can be represented with arbitrary precision if you add enough epicycles. For instance, here’s how a planet with an orbit shaped like an elephant might be modeled with 100 epicycles:⁷Credit to user anderstood on Mathematica StackExchange

But at that point, epicycles are horribly convoluted compared to the straightforward description that all planets go around the sun in ellipses (or in elephant-shapes). They work, but they aren’t very good, and the fact that they need so many parameters should be a tip that they’re not a good match for reality.

And this is the problem I have with this book and with a lot of similar theology: it feels too much like trying to rescue doomed theories. With enough special-purpose contraptions (in this case, the profound need for free will, or the need for pain specifically, or accepting that people always suffer for a reason), you can probably make a case that your theory is compatible with any data. So yeah, it’s possible to develop theodicies. But none of the ones I’ve heard of survive a little Bayesian reasoning.

References

• Tooley, Michael, “The Problem of Evil”, The Stanford Encyclopedia of Philosophy (Spring 2019 Edition), Edward N. Zalta (ed.), URL = <https://plato.stanford.edu/archives/spr2019/entries/evil/>.
• O’Connor, Timothy and Christopher Franklin, “Free Will”, The Stanford Encyclopedia of Philosophy (Spring 2021 Edition), Edward N. Zalta (ed.), forthcoming URL = <https://plato.stanford.edu/archives/spr2021/entries/freewill/>.
• Mayer, Jürgen, Khaled Khairy, and Jonathon Howard. “Drawing an elephant with four complex parameters.” American Journal of Physics 78.6 (2010): 648-649.