1weighing Evils: The C.S. Lewis Approach
1weighing Evils: The C.S. Lewis Approach
1weighing Evils: The C.S. Lewis Approach
LEWIS APPROACH
Introduction
The intuition that the vast amount of evil in our world counts against traditional theism in
a way that lesser quantities do not is prevalent both in lay circles and in philosophical treatments
of the problem of evil. For example, Hume says in the mouth of Demea: “Were a stranger to
drop, on a sudden, into this world, I would show him, as a specimen of its ills, a hospital full of
diseases, a prison crowded with malefactors and debtors, a field of battle strewed with carcasses,
a fleet foundering in the ocean, a nation languishing under tyranny, famine or pestilence.”1
evils involving small children, Ivan remarks, “Listen to me: I took children only so as to make it
more obvious. About all the other human tears that have soaked the whole earth through, from
crust to core, I don’t say a word . . .” [emphasis added].2 Both Demea and Ivan imply that a large
quantity of evil is a greater threat to the existence of God than significantly fewer evils would be.
Indeed, contemporary philosophers often take it for granted that the problem of evil is worse due
to the great quantity of evil in our world. For example, William Rowe says, “In light of our
experience and knowledge of the variety and scale of human and animal suffering in our world,
the idea that none of this suffering could have been prevented by an omnipotent being without
thereby losing a greater good or permitting an evil at least as bad seems an extraordinarily absurd
1 David Hume, Dialogues Concerning Natural Religion (London: Penguin Books, 1990), 106.
2 Fyodor Dostoevsky, The Brothers Karamazov, trans. by Richard Pevear and Larissa Volokhonsky (New
quantity of evil magnifies the problem of evil: “. . . even if there is a reason that would justify
God in permitting some evil, even a great deal of it, there may be no reason at all that would
Why would a greater quantity of evil make the existence of God less likely than a lesser
quantity? Presumably any quantity of evil is a problem for the existence of an omnipotent and
omnibenevolent deity, but contemporary objectors usually admit that some evils (or, at least,
some apparent evils) might be such that an omnipotent and omnibenevolent being would have a
morally sufficient reason to permit them. That is to say, they might not be gratuitous.5 However,
the greater the quantity of evil, the more likely it is that at least some of those evils are
gratuitous, and hence, the less likely it is that the God of traditional theism exists.
33 William Rowe, “The Problem of Evil and Some Varieties of Atheism,” in The Problem of Evil, ed. by
Marilyn McCord Adams and Robert Merrihew Adams (Oxford: Oxford University Press, 1990), 131; first published
44 “God, Evil, and Suffering,” in Reason for the Hope Within, ed. by Michael J. Murray (Grand Rapids, MI:
Eerdmans, 1999), 101. Howard-Snyder then delineates the argument from amount as follows: Premise [1]: There is
no reason that would justify God in permitting so much evil rather than a lot less. Premise [2]: If God exists, then
there must be such a reason. Conclusion: So, God does not exist, p. 102. After stating the argument, he concedes
that “It [argument from amount] is . . . the argument that I have found most troubling” (pp. 102-103).
55 Rowe himself concedes this: “Intense human or animal suffering is in itself bad, an evil, even though it
may sometimes be justified by virtue of being a part of, or leading to, some good which is unobtainable without it. . .
. While remaining an evil in itself, [it] is, nevertheless, an evil which someone might be morally justified in
determine the total quantity of evil in a world. Often evil is quantified on the model of the
hedonic calculus, the way in which pain is quantified in classical utilitarianism. On this method,
there are at least three scales of measure: the severity of the pain, the duration of the pain, and
the number of instances of pain. According to this calculus, each scale is commensurate with the
others, so the total quantity of pain in the world can be summed up. For example, one person
may have a pain twice as severe but equal in duration to the pain of two others, in which case the
first person’s pain would be 2n, and the other two persons’ pain would be 1n + 1n, resulting in
the same total amount of pain. We will call this approach to quantifying evil the additive
approach (AA). Possibly evils other than pains can be quantified by AA, but we will focus on the
problem of pain because it has the best chance of being quantifiable in this way.6 We will return
to the problem of separating the problem of pain from other evils and countervailing goods at the
Using AA as the method for measuring quantity of pain, the quantitative argument from
(1) The quantity of pain in the world should be measured by AA, and on that measure
the quantity is enormous.
(2) The greater the total quantity of pain, the more likely it is that some of the pain is
gratuitous.
(3) The more likely it is that there is gratuitous pain, the less likely it is that the God
6 Using this approach to sum moral evils is much more difficult. Does it make sense to say that one act of
injustice or betrayal or deception is twice as bad or forty times as bad as another? Is there a scale of duration for an
act of injustice or betrayal, or does that scale apply only to the pain and suffering of its effects?
of traditional theism exists.
In this paper we will examine C. S. Lewis’ objection to (1). We will argue that Lewis is
right to reject AA, but Lewis’ own position leads to some implausible judgments of quantity.
However, Lewis may still be right in his implied rejection of the above argument because even
though there is a sense in which he underestimates the quantity of pain, the quantity of pain in
that sense does not significantly increase the probability that some pain is gratuitous.
In a collection of his posthumous writings, Wittgenstein says, “The whole earth cannot be
in greater distress than one soul.”7 A similar sentiment is expressed in Chapter Seven of The
We must never make the problem of pain worse than it is by vague talk about the
“unimaginable sum of human misery.” Suppose that I have a toothache of intensity x: and
suppose that you, who are seated beside me, also begin to have a toothache of intensity x.
You may, if you choose, say that the total amount of pain in the room is now 2x: search
all time and all space and you will not find that composite pain in anyone’s
consciousness. There is no such thing as a sum of suffering, for no one suffers it. When
we have reached the maximum that a single person can suffer, we have, no doubt,
reached something very horrible, but we have reached all the suffering there ever can be
in the universe. The addition of a million fellow-sufferers adds no more pain. [emphasis
added]8
77 Wittgenstein, Culture and Value: A Selection from the Posthumous Remains, edited by G.H. von Wright,
88 C.S. Lewis, The Problem of Pain (New York: Touchstone, 1996), 103. Lewis’s comments here are
fascinating, but we find them in embryonic form. He leaves them without further explication, explication that might
have strengthened his case against AA. It is helpful to remember that Lewis’s aim was not to write a technical piece
of human misery,” and appealing to such a sum with anti-theistic motives—that this
incomprehensible sum jeopardizes God’s existence in a way that a much smaller sum does not.
The argument is grounded in a claim about the relationship between quantity and the limitations
of the conscious experience of pain. Summation of pain makes sense only within the
consciousness of a single being, argues Lewis, but no one conscious being can experience the
summation of all the pain in the world. In fact, no one being can experience the pain of more
than one conscious being. We cannot peer into anyone’s consciousness and there discover the
aggregate of pain of even two persons, much less all the pain in the universe. In speaking of the
“unimaginable sum” of suffering, objectors attempt to refer to an experiential evil that no one
actually experiences.
If pain can only be summed within the consciousness of a single individual, Lewis has an
argument that there is a maximum amount of pain in the world, and that is the maximum pain a
single conscious being can suffer. It does seem reasonable that given the neurological systems of
creatures, there is some amount of pain that is the most such a being can suffer at any one time,
and given the finitude of life, there is some amount of pain that is the most such a being can
suffer during an entire lifetime. It is less plausible to claim that this amount is the maximum
amount of pain in the world, and it is also implausible to think that there is no sense in which the
pain of many creatures is worse than the pain of one. Nonetheless, we think there is something
Let’s look at how Lewis’ approach to quantifying pain (we’ll call it “LA”) contrasts with
AA. For convenience, we will call a unit of pain a turp.9 Suppose A and B each suffer one turp of
pain in some situation s. According to AA, we can consider the experiences of A and B jointly
(b) s makes the existence of God less likely than a situation in which A suffers one
turp and B does not suffer.
LA, in contrast, sums the pain in s as one turp, contra (a), and concludes that s is less
threatening to the existence of God than the defender of AA acknowledges. Notice that Lewis
uses the rejection of the way pain in s is quantified by AA as a reason to reject (b).
We think both AA and LA are partly right, but to see the problems of each approach, we
need to compare a number of worlds. Assume that the maximum amount of pain a single
According to AA, the amount of pain in W2 equals the amount of pain in W3, and the pain in W4
99 Plantinga uses “turp” (turpitude = an instance of baseness, vileness, or depravity) to designate one unit of
evil, whereas we are using it to designate one unit of pain. Clearly, there are objections to the idea that there are
units of pain that have application to different sentient beings, and Lewis at least hints that he is sympathetic with
those objections, but we are granting the idea for the sake of argument.
equals the pain in W5. W2 and W3 have ten thousand times as much pain as W1, and W4 and W5
have a billion times as much pain as W1. According to LA, in contrast, the pain in W1,W2, and
W4 are equal, and the pain in W3 is the worst possible. World W5 is an impossible world.
We agree with Lewis that W3 is worse than W2, and W5 is impossible, given the
assumption that the maximum amount of pain a single individual can experience is well below a
billion turps.10 We think that the position that W3 is worse than W2 is prima facie justified, but
even if they are equal in quantity of pain, as AA maintains, W3 needs a greater justifying reason
for its existence than W2. This is evident if God’s properties include the desire to have a
relationship with his creatures, or at least, with each creature capable of having a relationship.
Surely in that case God would need a much more serious reason to permit a world in which one
of his creatures suffers the greatest amount of pain of which that being is capable rather than a
world in which the same number of turps of pain is distributed over ten thousand creatures. So
whether W2 and W3 are compared by LA or by AA, we think that Lewis is right in his implied
But LA leads to counterintuitive results also. Surely, worlds W1,W2, and W4 are not
actually equal in total amount of pain, or at least, the pain in W2 , is greater in some sense than
the pain in W1, and the pain in W4 is greater still. That is implied by the fact that we believe it is
praiseworthy for a person to choose to have her hand burned to spare ninety-nine others from
such pain. Presumably we think that her pain prevents pain that is greater in some sense. Her act
is admirable, not only because it is done for others, but because it prevents a greater degree of
10
1 Of course, we are merely stipulating the numerical value of the maximum amount of pain a single
conscious being can experience, but if there is a maximum amount, there will be some number n which is such that
it is impossible that there is a world containing an individual experiencing pain greater than n.
pain from happening.
This must be the case even on Lewis’ assumption that pain can only be summed within
the consciousness of a single individual. If the total pain in one person’s consciousness is a bad
thing, then the total pain in the consciousness of each of a thousand or a billion individuals is
each a bad thing, and God needs a reason to permit each of those bad things, even if the
consciousness of each individual is taken individually and the aggregate cannot be summed.
Another way to put this point is that there are as many problems of pain as there are sentient
beings. W2 has ten thousand problems of pain and W4 has a billion, whereas W1 has only one.
This is the sense in which W1,W2, and W4 are not equal in the problem of pain even if they are
equal in total quantity of pain, as Lewis maintains. The suffering in W2 and W4 brings with it an
increased burden on God to rectify or justify the suffering. Indeed, he needs ten thousand reasons
How does the fact that there are many problems of pain affect the probability that pain is
gratuitous, premise (2) of the quantitative argument from evil as we have stated it? Let’s
compare W1,W2, and W3. We suggested above that it is more difficult to see how God could have
a reason to permit W3 than to permit W2 even if the sum of pain is the same in the two worlds.
The inference that the pain in W3 is gratuitous is stronger than the parallel inference for W2. We
want to argue now that W2 may be no more difficult to justify than W1, or at any rate, not
significantly harder to justify, given relevantly similar conditions for the individuals in W2. The
inference to the probable gratuitousness of the pain in W2 is comparable to the inference to the
To see this, consider what we normally say about sufficient justifying reasons that are
repeatable in similar situations. If one person has a reason to swallow her bitter-tasting medicine
to cure her infection, any person has a reason to take the medicine, given relevantly similar
conditions. If one person has a reason to have a cancerous tumor removed despite the pain it will
cause, so do ten thousand or a million people with cancerous tumors in similar circumstances. If
one person has a reason to ask forgiveness of a person she loves whom she has harmed, so does
This point is generalizable. Any time a person A has a sufficient justifying reason to
permit p in situation s, anybody has a sufficient reason to permit something relevantly similar to
but the principle is both fair and widely accepted. In fact, it is probably universally accepted. In
any case, the principle need not be embraced in its full generality for it to be applicable to God’s
permission of evil. We are appealing to the following more limited principle, which we can call
We can see the expectation of PRR in the following example. If a worker comes down
with the flu, this is a legitimate reason for him to stay home from work. And if sickness is a
legitimate reason to stay home from work on this occasion, then it retains its legitimacy on other
occasions. If he stays home from work a year later because of sickness, his boss cannot
legitimately say, “You need a new reason this time.” If sickness is a legitimate reason on one
occasion, it is on another occasion, assuming that the conditions are relevantly similar.
clear. If God’s reasons to permit one person’s pain are relevantly similar to his reasons for
permitting another person’s pain, W2 and W4 are as justifiable as W1, assuming that the pains of
circumstances. God does not have to have a collective reason for permitting the pains in W2 or
W4. It is sufficient if God has repeatable reasons for permitting the pain of each individual.
Given the view that pain is additive, premise (2) of the quantitative argument from evil
above has another problem. The probability that a given quantity of pain is gratuitous cannot be
separated from other factors that provide the context in which the pain occurs and through which
the pain must be understood. To see this, consider a number of worlds in which individuals
suffer the pain of a needle prick. Assume that one turp = the pain of one needle prick.
Let W6 be a world in which a single person suffers one hundred turps and is successfully
inoculated against a fatal but painless disease.
W7 consists of a single person who suffers one hundred turps and is not successfully
inoculated against the fatal but painless disease.
W8 consists of one thousand persons each suffering one turp and each is successfully
inoculated against the fatal but painless disease.
W9 consists of one thousand persons each suffering one turp and each is not successfully
inoculated against the fatal but painless disease.
According to AA, W6 and W7 contain the same amount of pain, as do W8 and W9, and the
latter pair are ten times as bad as the former pair. But surely the problem of pain is not equal in
W6 and W7, nor is it equal in W8 and W9. Other things being equal, W7 is worse than W6 and W9
is worse than W8. W7 needs a greater justifying reason than W6 and W9 needs a greater justifying
reason than W8. Furthermore, the pain in W8 is less likely to be gratuitous than the pain in W7
even though the former is considerably worse on the additive approach. And if we were right in
We have argued that Lewis’ point that the pains of multiple individuals is not additive is
plausible, and for that reason W3 is worse than W2. But there is still a sense in which a greater
number of individuals suffering pain increases the quantity of evil. That is because even on
Lewis’ assumption that pain is only additive within the consciousness of each individual, there
are as many problems of pain as there are suffering individuals. However, a world with multiple
individuals suffering pain has a different relation to the probability that the pains are gratuitous
than a world with the same total quantity of pain, but where the pain is suffered by a single
individual. The pain suffered by a single person needs a justifying reason that can explain the
quantity of pain suffered by that individual. Pain suffered by multiple individuals in the same
circumstances can often be justified by a reason that explains any one of them. The probability
that the pain in W2, or W4 is gratuitous may be no greater than the probability that the pain in W1
is gratuitous, and is far less than the probability that the pain in W3 is gratuitous. W5 cannot have
gratuitous pain since it is not a possible world. Furthermore, the probability that a given quantity
of pain is gratuitous cannot be calculated apart from the connection between the pain and
countervailing goods that are not additive, such as protection from a fatal but painless disease.
We conclude that both premises (1) and (2) of the quantitative argument from evil are probably
false.
We are not suggesting that the problem of pain is not severe. It follows from our
reasoning that a world containing creatures that suffer the maximum amount of pain of which
those creatures are capable has problems of pain that threaten the existence of God. There may
well be creatures in our world who suffer the maximum amount of pain for some period of their
lives, and there needs to be an explanation for their suffering. We may think of Rowe’s suffering
fawn in the raging forest fire, twelve year old Ashley Jones, brutally beaten, raped, and
murdered, and the cement truck driver who accidentally ran over his three year old daughter
while she played in the front yard. These are appalling cases of suffering and evil that bring us
almost unbearable emotional angst and cry out for explanation, but we do not think that adding
more suffering creatures to the world significantly reduces the probability that there is a God if
the suffering is in similar situations. Given PPR, we think that the reason for one can be applied
to others. We therefore need not embrace the claim of Hume, Dostoyevsky, Rowe, Howard-
Snyder, and many other philosophers that the quantity of evil suffered by many different