. 3
( 8)


God and the Reach of Reason

the relation between God and the moral law to be represented?”26
Before providing his own answer to this question, Lewis rejects two
other possible answers, declaring that “God neither obeys nor cre-
ates the moral law. The good is uncreated; it never could have been
otherwise; it has in it no shadow of contingency; it lies, as Plato
said, on the other side of existence.”27 This passage is important for
two reasons. First, it helps us determine Lewis™s own view on the
relationship between God and the moral law by telling us what, in
Lewis™s eyes, that relationship is not. Second, it indicates that in this
context Lewis uses “the good” and “the moral law” interchangeably,
a piece of information that is crucial in deciphering Lewis™s position
here. That position is stated in the following tricky passage:

[W]hat lies beyond existence, what admits no contingency, what lends
divinity to all else, what is the ground of all existence, is not simply
a law but also a begetting love, a love begotten. . . . God is not merely
good, but goodness; goodness is not merely divine, but God.28

Commenting on this passage in his doctoral dissertation on Lewis™s
philosophical writings, Steve Lovell interprets Lewis as claiming that
“God = (His?) Goodness” and remarks: “I must confess to not having
much idea about what [this thesis] could mean. It appears to assert an
identity relation between God and an abstract object, indeed between
God and a property.”29 I too have trouble grasping the claim that a
mindlike Higher Power is identical to the property of goodness.30
I am not sure, however, that Lovell™s interpretation of the crucial
passage is quite right. In light of my suggestion that in this context
the good = the moral law, a more likely interpretation is that Lewis is
suggesting that God is identical to the moral law. This interpretation
makes sense of the discussion as a whole: Lewis begins with the
question of how God is related to the moral law, says that God neither
obeys it nor creates it, and concludes that God simply is the moral law.
If this proposal makes sense, it gives Lewis a way of rejecting (RC)
by suggesting a way God might be good other than by conforming
to an independent moral law:

(LA1) Being identical to the moral law is a way of being good.
Beyond Nature

I ¬nd the proposal that God is the moral law roughly as puzzling as
the proposal that God is the property of goodness. Insofar as I can
understand the claim that God is the moral law, it seems to be the
claim that God is identical to a conjunction of ethical facts, facts like:
It is morally wrong to torture the innocent just for fun. It is hard to see
how a conjunction of such facts could be a mindlike Higher Power,
much less the personal God of Christianity. The obscurity of Lewis™s
proposal here seems to me to be a serious strike against it; however,
I will leave the proposal on the table for the sake of examining its
implications for Lewis™s moral argument.31 Before considering such
implications we must examine two other accounts of divine goodness
that I think are suggested by Lewis™s writing.
In Mere Christianity, Lewis appears to conceive of the relationship
between God and the moral law in a way that is incompatible with
the position he takes in “The Poison of Subjectivism.” For instance,
in Mere Christianity he says that “there is a real Moral Law, and a
Power behind the law.”32 The language here suggests that the Moral
Law and the Higher Power are distinct, and a natural interpretation
of the overall discussion in Mere Christianity is that the Higher Power
is the author or creator of the Moral Law, one of the two options
explicitly rejected in “The Poison of Subjectivism.”
This suggests that in Mere Christianity Lewis is working with a con-
ception of divine goodness distinct from the one he proposes in “The
Poison of Subjectivism.” But what is this other conception? The fol-
lowing passage sheds some light: “God is quite de¬nitely ˜good™ or
˜righteous™, a God who takes sides, who loves love and hates hatred,
who wants us to behave in one way and not in another.”33 We
should also recall Lewis™s remark that the Higher Power is “intensely
interested . . . in fair play, unsel¬shness, courage, good faith, honesty
and truthfulness.”34 These passages suggest the view that one way
of being good is by loving certain things (and perhaps hating cer-
tain other things).35 If this is right, we have another alternative to

(LA2) Loving love, fair play, unsel¬shness, courage, good faith,
honesty, and truthfulness is a way of being good.
God and the Reach of Reason

A third way of thinking of divine goodness is the conception of divine
goodness that Lewis employs in The Problem of Pain:

(LA3) Desiring that human beings attain genuine happiness (that
they freely love God and strive to become Christlike) is a way
of being good.

At this point I would like to review the discussion so far and say a
bit about what is to come. We started with an exposition of Lewis™s
moral argument from Mere Christianity. Against that kind of argu-
ment, Russell raises a puzzle about the nature of God™s goodness.
Russell™s puzzle appears to rely on (RC), the claim that there is only
one way a being can be good: by conforming to a moral law not
authored by the being in question. Russell and Lewis rightly agree
that such a view makes it impossible for God to be good and to be
the author of the moral law. Russell concludes from this that moral
arguments like Lewis™s fail. Lewis, on the other hand, concludes that
(RC) is false. I have suggested that Lewis™s writings suggest three
alternatives to (RC): God can be good by being identical to the moral
law (LA1), by loving certain things (LA2), or by desiring that humans
attain genuine happiness (LA3).36 The next order of business is to
examine the implications for Lewis™s moral argument of each of the
In Mere Christianity, Lewis not only appeals to morality to argue
for the existence of a Higher Power; he also argues against a view he
calls Dualism. I think it is no mere coincidence that Lewis undertakes
both projects in the same work. I believe that if Dualism is a tenable
view, this fact constitutes the basis of a powerful objection to Lewis™s
moral argument. I will argue that if either (LA2) or (LA3) is true,
then Lewis fails to refute Dualism, and hence his moral argument
fails. If this is right, it means that Lewis™s moral argument can suc-
ceed only if (LA2) and (LA3) are false. On the other hand, if (LA1)
is true, then Lewis™s objection to Dualism goes through. However,
(LA1) leads to two other problems for Lewis™s moral argument. The
¬rst is that the argument depends on an account of the relationship
between God and the moral law that is at best obscure and at worst
incoherent. The second is that (LA1) points in the direction of a
Beyond Nature

plausible atheistic explanation of Lewisian moral phenomena. The
existence of such an explanation renders Lewis™s moral argument
toothless. If all of this is correct, the upshot is that Lewis has no under-
standing of the nature of divine goodness that enables his moral
argument to succeed “ and, to the extent that the atheistic account
of Lewisian moral phenomena that I will sketch is plausible, Lewis™s
argument fails in any case. I turn now to the task of making this
Let us begin with Lewis™s case against Dualism. My main goal in
the upcoming section is to establish two claims: ¬rst, that the truth of
either (LA2) or (LA3) ruins Lewis™s argument against Dualism, and
second, that the failure of the argument against Dualism entails the
failure of the moral argument as well.

2.2.3 Lewis™s Attack on Dualism
The view that Lewis calls Dualism is common to a number of re-
ligious alternatives to Christianity, including Zoroastrianism and
Manicheanism.37 It is the view that “there are two equal and inde-
pendent powers at the back of everything, one of them good and the
other bad, and that this universe is the battle¬eld in which they ¬ght
out an endless war.”38 Zoroastrianism is an ancient Persian religion
that still has adherents in Iran and southern Asia. Manicheanism,
on the other hand, seems to have pretty much died out, but it was
at one time one of the main rivals of Christianity, particularly in the
early days of Christianity. Saint Augustine adhered to the view for
some time, and in the thirteenth century Pope Innocent III launched
a crusade against the Cathars in southern France, who held a version
of Manicheanism.39 In Mere Christianity, Lewis describes Dualism as
“next to Christianity . . . the manliest and most sensible creed on the
market.”40 In a 1942 letter, however, he offers a more blunt explana-
tion for taking on Dualism: “You wouldn™t be surprised at the space
I give to Dualism if you knew how attractive it is to some simple
Despite its manliness and attractiveness to simple minds, the view
has dif¬culties, says Lewis. One alleged problem stems from the claim
God and the Reach of Reason

that the two Highest Powers have opposite moral qualities, one being
good, the other evil. This claim entails that there is

a third thing in addition to the two Powers: some law or standard or
rule of good which one of the powers conforms to and the other fails
to conform to. But since the two powers are judged by this standard,
then this standard, or the Being who made this standard, is farther
back and higher up than either of them, and He will be the real God.42

The crucial premise here is that the existence of equal, independent
Powers, one good, one evil, entails the existence of a third Power
superior to both of the ¬rst two. Since Dualism is incompatible with
the existence of such a third Power, Lewis aims to show that Dualism
is internally inconsistent.
The truth of either (LA2) or (LA3) seems to undermine this crucial
premise. Let us ¬rst consider the implications of (LA2) in this regard:

(LA2) Loving love, fair play, unsel¬shness, courage, good faith,
honesty, and truthfulness is a way of being good.

Given (LA2), it is natural to suppose that hating love, fair play, and
so on is a way of being evil. But if this is right, then it is hard to see
why Dualism would require a third Higher Power. The existence of
two equal and opposite Powers, one of which loves love, fair play,
and the rest, the other of which hates these things, without a third
Higher Power superior to these two, seems perfectly coherent. Sim-
ilar considerations apply to Lewis™s third alternative to (RC):

(LA3) Desiring that human beings attain genuine happiness (that
they freely love God and strive to become Christlike) is a way
of being good.

Like (LA2), (LA3) suggests a corresponding account of evil “ in this
case, that desiring that human beings fail to attain (or attain the
opposite of) genuine happiness is a way of being evil. And, again,
it is hard to see why, under this understanding of good and evil,
the existence of a good Power and an evil Power would require the
existence of a third Even Higher Power.
Beyond Nature

Finally, consider the principle suggested by Lewis™s discussion in
“The Poison of Subjectivism”:

(LA1) Being identical to the moral law is a way of being good.

Suppose that this is true and that there are only two ways of being
good “ either (i) by obeying a moral law not of one™s own creation
or (ii) by being identical to the moral law. It follows that the good
Power of Dualism is either subject to a moral law it did not create or
is itself the moral law. Under the ¬rst alternative, Lewis™s claim that
we must posit a third Higher Power superior to the other two Powers
seems reasonable; in any case, I will assume that the Dualist cannot
consistently say that the good Power of Dualism is subject to a moral
law it did not create. This means that the Dualist must identify the
good Higher Power with the moral law. This in turn implies that the
evil Power is subordinate to the good Power, since presumably what
makes the evil Power evil is that it violates a moral law to which it is
subject. But since we are supposing that the good Power is the moral
law, it follows that the evil Power is subject to the good Power “
and that conclusion also con¬‚icts with Dualism. Therefore, given
the supposition put forth at the start of this paragraph, Lewis™s ¬rst
argument against Dualism goes through, but if either (LA2) or (LA3)
is true, the argument fails. Thus we may (tentatively) conclude that
Lewis™s objection to Dualism succeeds only if both (LA2) and (LA3)
are false.
However, matters are complicated by the fact that Lewis presents a
second objection to Dualism in Mere Christianity. If this other objection
succeeds, then the limitations of the ¬rst objection to Dualism are
irrelevant. To support my contention that Lewis can refute Dualism
only if (LA2) and (LA3) are false, I will make the case that this second
objection to Dualism simply fails. Here is the essence of the second

If Dualism is true, then the bad Power must be a being who likes
badness for its own sake. But in reality we have no experience of
anyone liking badness just because it is bad. . . . [W]ickedness, when
you examine it, turns out to be the pursuit of some good in the wrong

God and the Reach of Reason

way. You can be good for the mere sake of goodness: you cannot be
bad for the mere sake of badness.43

The central premise of this objection is that loving evil for its own
sake is impossible. Yet Dualism requires that this be possible. Hence,
Dualism is false. To support the central premise, Lewis appeals to
experience: We never encounter people who love evil for its own
This argument is unconvincing. One problem is that from the fact
that we have no experience of beings who love evil for its own sake it
hardly follows that such beings are impossible.44 Indeed, it is some-
what surprising to ¬nd this sort of argument coming from some-
one like Lewis, who believes in supernatural, transcendent beings
remarkably unlike any we experience. Furthermore, it is far from
clear that beings who love evil for its own sake have not actually
existed. For instance, if we take Saint Augustine at his word, the
youthful Augustine was just such a being:

I stole things which I already had in plenty and of better quality. Nor
had I any desire to enjoy the things I stole, but only the stealing of them
and the sin. . . . Such was my heart, O God, such was my heart. . . . Let
that heart now tell You what it sought when I was thus evil for no
object, having no cause for wrongdoing save my wrongness. The mal-
ice of the act was base and I loved it “ that is to say I loved my own
undoing, I loved the evil in me “ not the thing for which I did the
evil, simply the evil: my soul was depraved, and hurled itself down
from security in You into utter destruction, seeking no pro¬t from
wickedness but only to be wicked.45

Notice in particular Augustine™s remark that he loved “not the thing
for which [he] did the evil” but loved “simply the evil.” Surely the most
natural interpretation of this remark is that Augustine is claiming to
have loved evil for its own sake “ precisely the thing that Lewis claims
is impossible. Intriguingly, the younger, atheistic Lewis maintained
that the kind of desire Augustine describes is not only possible but
widespread. In a 1923 diary entry Lewis endorses the view that “most
of us could ¬nd positive Satanic badness down there somewhere, the
desire for evil not because it was pleasant but because it was evil.”46
Beyond Nature

If the psychological claims of Augustine and young Lewis are correct,
then the elder Lewis™s argument fails. At the very least, it is hard to
resist the conclusion that the kind of motivation Augustine describes
is possible, and this seems to be enough to save Dualism from Lewis™s
Against this criticism of Lewis™s argument it might be suggested
that even though Lewis appeals to experience to establish his crucial
premise, an a priori argument (one that does not rely on experience)
for the same conclusion is available. That argument runs as follows:
Every action must aim at a goal that the agent regards as good or
worthwhile in some respect. But no goal can be regarded as good or
worthwhile merely on the grounds that it is wicked or evil; hence,
no action can be directed at evil purely for its own sake. With respect
to the case of Augustine, a defender of this argument might suggest
an alternative interpretation of Augustine™s remarks. This alternative
interpretation is based on the fact that Augustine says that he got a
pleasurable thrill from doing the things he knew he should not do.
It was at least partially for the sake of this pleasure that he did these
things, and pleasure, after all, is a good “ a good that Augustine
pursued in ways he should not have.47
My response to this line of reasoning is that the question of what
kinds of considerations can motivate agents is an empirical question
about human psychology. The a priori argument simply assumes a
certain answer to this question “ that agents can pursue only goals
that they take to be good. But I do not see why such a claim should be
accepted a priori.48 It is true that Augustine says that he experienced
pleasure as a result of performing evil actions. But it does not follow
from this that he performed the wicked actions (even partially) for
the sake of that pleasure. In general, from the fact that action A had
a particular consequence C, it does not follow that the agent who per-
formed A did so for the sake of C. A simple example illustrates this
point: Every physical action I have performed has among its con-
sequences the displacement of some oxygen molecules, but I have
never once performed an action for the sake of displacing oxygen
molecules. That Augustine performed evil actions for the sake of
pleasure is one possibility; another possibility is that he performed
God and the Reach of Reason

evil actions purely because they were evil and that when he did so he
experienced a thrill as a consequence of doing evil for its own sake. The
case of Augustine presents prima facie evidence against the view that
agents cannot pursue evil for its own sake. The case can be construed
in such a way as to be consistent with this view, but such a construal
is not the only plausible one.
It seems to me that the proposed a priori argument begs the ques-
tion against the Dualist by assuming that evil cannot be pursued
simply because it is evil. Lewis himself tries to establish this premise
by relying on the empirical claim that people in fact never pursue
evil for it own sake. This empirical claim is questionable and, even
if true, fails to establish the crucial premise of Lewis™s argument. I
conclude that Lewis™s second objection to Dualism fails.
If everything I have said in this section so far is correct, then Lewis
can refute Dualism only by rejecting both (LA2) and (LA3). To refute
Dualism, he must insist that there are just two ways a being can be
good “ either by being the moral law, or by obeying a moral law the
being did not create. The importance of this result lies in the fact
that if Lewis cannot rule out Dualism, then his moral argument is in
trouble, as I shall now argue.
Consider the Apostle Paul™s famous remarks on his own efforts to
do the right thing: “I can will what is right, but I cannot do it. For
I do not do the good I want, but the evil I do not want is what I
do.”49 Charles Freeman remarks: “No one reading Paul can ignore
the powerful emotional force of this message: human beings live at
the centre of a cosmic drama that reaches to the core of each person-
ality as the forces of good and evil battle within the individual.”50 We
¬nd within ourselves not just promptings inclining us toward good-
ness, but also promptings inclining us toward evil, and the con¬‚ict
between the two can produce internal turmoil. Indeed, this inner
struggle is one of the central features of human moral experience
and has been discussed by every important moral philosopher in the
Western tradition. This brief remark by Immanuel Kant captures the
phenomenon nicely: “Be a man ever so virtuous, there are in him
promptings of evil, and he must constantly contend with these.”51 Is
not Dualism a perfectly reasonable explanation of this phenomenon?
Beyond Nature

Among the data Lewis draws on in developing his moral argument
are certain moral emotions, including guilt and a sense of obligation.
Lewis suggests that these phenomena point toward a good Higher
Power. But we also ¬nd within ourselves temptations and evil incli-
nations. If it is reasonable to suppose that positive moral emotions
are indicative of a good Higher Power, is it not equally reasonable
to suppose that negative moral emotions are indicative of an evil
Higher Power?
Lewis notes that traditional Christianity includes the belief that
there is a “Dark Power” who was “created by God, and was good
when he was created, and went wrong.”52 This Dark Power is “be-
hind death and disease, and sin.”53 Perhaps our internal struggle mir-
rors the struggle between the two Powers of Dualism. These consid-
erations suggest an argument for Dualism that is structurally parallel
to Lewis™s moral argument:

A Moral Argument for Dualism

1. Positive and negative moral phenomena exist.
2. The best explanation of the existence of such moral phenomena
is the existence of a Higher Power (or Powers) that created the
3. So: There is a Higher Power (or Powers) that created the universe
(from 1 and 2).
4. The Higher Power(s) issue instructions and want us to engage in
morally right conduct but also tempt us and want us to engage
in morally wrong conduct.
5. If (4), then there are two Higher Powers that created the uni-
verse, one good, one evil.
6. Therefore, there are two Higher Powers that created the uni-
verse, one good, one evil (from 4 and 5).

We have, then, parallel arguments for incompatible conclusions.
Unless there is an independent reason to prefer one argument to the
other, we arrive at the following stand-off: It is reasonable to endorse
one argument only if it is also reasonable to endorse the other.
But it is not reasonable to endorse both; hence, it is not reasonable
God and the Reach of Reason

to endorse either. Of course, Dualism is not a live option for most
people, but this is beside the point. The issue at hand is whether it is
reasonable to believe in a single good Higher Power on the basis of
Lewis™s moral argument; the fact that there are few Dualists around
today does not constitute a good reason to prefer Lewis™s argument
to the moral argument for Dualism. If Dualism could be refuted
directly, this of course would break the stand-off. I suspect that this
is at least part of the reason why Lewis attempts such a refutation
in Mere Christianity. Without the refutation, Lewis has not given us a
good reason to prefer his moral argument to the moral argument for
In this section I have argued for two main claims. First, I™ve argued
that Lewis™s attack on Dualism in Mere Christianity succeeds only if
there are only two ways a being can be good “ either by being the
moral law, or by obeying a moral law the being did not create. Second,
I™ve argued that Lewis™s moral argument succeeds only if Lewis™s
attack on Dualism succeeds. It follows from these two claims that
Lewis™s moral argument succeeds only if there are just two ways a
being can be good “ either by being the moral law, or by obeying a
moral law the being did not create. Given the obscurity and possible
incoherence of the notion that God is the moral law, this result makes
serious trouble for Lewis™s moral argument. In the next section I will
examine a different kind of objection to Lewis™s moral argument. The
essence of this other objection is that Lewisian moral phenomena are
perfectly at home in a universe devoid of Higher Powers altogether.

2.2.4 Godless Objective Morality
Here again is Lewis™s moral argument:

Lewis™s Moral Argument

1. Lewisian moral phenomena exist.
2. The best explanation of the existence of Lewisian moral phenom-
ena is the existence of a Higher Power that created the universe.
3. So: There is a Higher Power that created the universe (from 1
and 2).
Beyond Nature

4. The Higher Power issues instructions and wants us to engage in
morally right conduct.
5. If (4), then there is a good, mindlike Higher Power that created
the universe.
6. Therefore, there is a good, mindlike Higher Power that created
the universe (from 4 and 5).

I suggested earlier that Lewis™s argument relies on an inference to
the best explanation. This suggests two strategies for attacking the
argument. One strategy is to deny the existence of Lewisian moral
phenomena and reject the ¬rst premise. For example, some resist
arguments like Lewis™s moral argument by denying the existence of
objective morality altogether. I will not consider this approach here
for the simple reason that I ¬nd it implausible. A second strategy
is to ¬nd explanations for Lewisian moral phenomena that do not
involve Higher Powers and reject the second premise. Both Russell
and Hume pursue the second strategy with respect to at least some
of the Lewisian moral phenomena.
Let us remind ourselves of the nature of Lewisian moral phenom-
ena. There are three such phenomena: (i) that humans have certain
moral obligations (moral facts), (ii) that most humans know what at
least some of these obligations are (moral knowledge), and (iii) that
most humans experience various emotions related to these obliga-
tions, such as guilt and a sense of obligation (moral emotions).
The faculty that is commonly thought to be responsible for com-
ponents (ii) and (iii) is conscience, which is typically thought to be
possessed by all or nearly all humans. Some passages indicate that
Lewis thinks that conscience has a divine origin. For example, in
Mere Christianity, he describes the Higher Power as something that
“appears in me as a law urging me to do right,” and he describes the
moral law as something “which He has put into our minds.”55 In The
Problem of Pain, he remarks that God “speaks in our conscience.”56
We have already considered Russell™s criticism of the moral argu-
ment from his 1927 essay “Why I Am Not a Christian.” Several years
later, Russell seemed to favor a different kind of response to the
moral argument. This other response is based on the alleged variation
God and the Reach of Reason

in the deliverances of conscience from one person to the next. For
instance, in his 1939 essay “The Existence and Nature of God,”
Russell writes:

[C]onscience varies with different people in different instances. . . .
Take such a thing as human sacri¬ce. It has existed in pretty nearly
all races. It is the normal phase of a certain stage in the development
of the race. To those who practiced it, it was an essential part of their
religion. . . . You will ¬nd that what your conscience tells you varies
according to the age and place. . . . 57

In 1948, the BBC broadcast a debate on the existence of God between
Russell and the Jesuit philosopher F. C. Copleston. In that debate
Copleston suggested that “the consciousness of moral law and obli-
gation [is] best explained through the hypothesis of . . . an author
of the moral law.”58 Russell™s immediate response was that “the
moral law . . . is always changing. At one period in the development
of the human race, almost everybody thought cannibalism was a
Let us grant Russell™s claim that the deliverances of conscience
vary widely from person to person. What, exactly, is the signi¬cance
of this supposed to be? Russell™s idea is that this variation is evidence
that conscience has not a divine origin but rather an earthly one:
“[C]onscience is the stored up discomfort due to disapproval expe-
rienced or imagined in the past, particularly in early youth. So far
from having a divine origin, it is a product of education, and can be
trained to approve or disapprove as educators see ¬t.”60 In the debate
with Copleston, Russell remarked that “the feeling that one has
about ˜ought™ is an echo of what has been told one by one™s parents
or one™s nurses.”61
Russell™s argument, then, seems to be that the variation in the
deliverances of conscience from one person to the next indicates
that conscience is entirely a product of the education (or, more pejo-
ratively, the conditioning) that one receives while very young. The
moral precepts ingrained during youth persist into adulthood, and
they “come up as if they had an external source and seem like the

Beyond Nature

voice of God,” but this is merely an illusion.62 In fact, which precepts
one is instilled with is often a matter of chance, since it is often a
matter of chance who is responsible for one™s education while one is
young, and consequently the precepts one ¬nds instilled in oneself
may turn out to be false. Russell had some personal experience with
this phenomenon, and this may have in¬‚uenced his views on the
issue. In his autobiography, Russell has this to say concerning his
¬rst wife, Alys:

She had been brought up, as American women always were in those
days, to think that sex was beastly, that all women hated it, and that
men™s brutal lusts were the chief obstacle to happiness in marriage.
She therefore thought that intercourse should only take place when
children were desired.63

Russell™s argument is not intended to show, nor does it in fact show,
that there are no universal, objective moral truths. The mere fact that
we can be indoctrinated with false beliefs in a given area does not
imply that there are no truths in this area; if it did, nearly all truths
would vanish. The point of the argument, rather, is to establish that
the explanation for our moral beliefs and emotional dispositions lies
not in divine activity but instead in our upbringing.
In re¬‚ecting on this argument, it is important to distinguish the
claim (a) that there is signi¬cant variation in the moral precepts
held by different people from the claim (b) that there are no moral
precepts that are held by everyone (or almost everyone). Russell
offers various examples of variation in the deliverances of conscience,
but of course such examples at best support the weaker claim (a),
whereas it seems that his argument requires the stronger claim (b).
From (a) the most that would follow is that conscience can be cor-
rupted through a bad upbringing, but it is hard to see how we can
reasonably move from this claim to the desired conclusion that con-
science is entirely a product of one™s upbringing. There is no indication
that Lewis maintains that conscience is entirely incorruptible or that
any old moral belief or feeling a person has is produced by that per-
son™s uncorrupted conscience. The presence of some moral beliefs

God and the Reach of Reason

and emotions that have an earthly origin is perfectly consistent with
the presence of others that have a divine origin.
It seems, therefore, that Russell™s argument fails to establish its
intended conclusion. The fundamental problem with the argument
is that establishing the presence of disagreement is not the same as
establishing the absence of agreement. Russell has accomplished at
most the ¬rst thing, but his argument requires that he accomplish
the second thing. Moreover, in the Appendix to The Abolition of Man,
Lewis presents a host of moral precepts that he claims to be univer-
sally (or nearly universally) recognized, together with textual evi-
dence from various traditions indicating recognition of the relevant
precept within each tradition.64 To give his argument legs, Russell
would need to discuss such alleged universal precepts and show that
they are not so widely held after all, and this is something that he
does not do.
Of course, even if there are moral precepts held by just about
everyone, it could be the case that such shared precepts are a result
of education; the common precepts could be a result of common ele-
ments of education. But Russell is trying to establish that conscience
is entirely a product of education, not merely that it might be.
Perhaps more problematic for Lewis™s argument than variation in
the deliverances of conscience is the fact that some people appar-
ently lack a conscience altogether. Psychopathy (sometimes called
“sociopathy”) is a personality disorder characterized by, among other
things, the absence of the capacity to experience various emotions,
including empathy, love, and guilt.65 Psychopaths know the differ-
ence between right and wrong in some sense; at least, they recognize
that other people view certain acts as being right or wrong and are
able to apply the terms “right” and “wrong” appropriately. But right-
ness and wrongness are of no signi¬cance to psychopaths, who liter-
ally do not care about morality. Robert Hare, a psychologist who has
studied psychopathy for over a quarter of a century, puts it this way:

[T]hey know the rules but follow only those they choose to fol-
low, no matter what the repercussions for others. They have lit-
tle resistance to temptation, and their transgressions elicit no guilt.

Beyond Nature

Without the shackles of a nagging conscience, they feel free to satisfy
their needs and wants and do whatever they think they can get away

Lewis acknowledges that there may be “an odd individual here and
there” who does not know the Law of Nature, “just as you ¬nd a
few people who are colour-blind or have no ear for a tune.”67 Inter-
estingly, those who study psychopaths have used similar analogies
to describe psychopathy. Hare quotes two researchers who declare
that when it comes to emotion, a psychopath “knows the words
but not the music.”68 Hare himself uses color-blindness to explain
psychopathy as follows:

The psychopath is like a color-blind person who sees the world in
shades of gray but who has learned how to function in a colored
world. He has learned that the light signal for “stop” is at the top of
the traf¬c light. When the color-blind person tells you he stopped
at the red light, he really means he stopped at the top light. . . . Like
the color-blind person, the psychopath lacks an important element
of experience “ in this case, emotional experience “ but may have
learned the words that others use to describe or mimic experiences
that he cannot really understand.69

Recall the “problem of not enough pain” discussed in the previous
chapter. It seems that the phenomenon of psychopathy may pose a
similar problem for Lewis™s view. The problem stems from Lewis™s
idea that the human conscience is a tool that God uses to commu-
nicate with us. More precisely, conscience is a tool that God uses
to get us to recognize our need for Him. Lewis says: “Christianity
tells people to repent and promises them forgiveness. It therefore
has nothing (as far as I know) to say to people who do not know
they have done anything to repent of and who do not feel that they
need any forgiveness.”70
Psychopaths are incapable of feeling that they need forgiveness.
Has God abandoned them? Psychologists estimate that about 4 per-
cent of human beings are psychopaths (at least in the West).71 In light
of Lewis™s views on divine goodness and human happiness described
in the previous chapter and his idea that conscience is an important
God and the Reach of Reason

tool that God uses to lead human beings to genuine happiness, what
are we to make of the fact that roughly one in twenty-¬ve human
beings is a psychopath? Recall the fourth premise of Lewis™s moral

4. The Higher Power issues instructions and wants us to engage in
morally right conduct.

The phenomenon of psychopathy seems to undermine this premise
to some extent. If the Higher Power wants us to engage in morally
right conduct, why does He permit so many of us to lack the emo-
tional equipment essential to doing so? I am not sure that this objec-
tion is decisive, primarily because of the possibility of a justi¬cation
for psychopathy that lies beyond our understanding, but it seems
to me that psychopathy joins non-victim-improving natural child
suffering as a phenomenon that does not ¬t very well with Lewis™s
overall view of things.
In any case, let us put aside psychopathy and focus on the vast
majority of human beings who do have the basic elements of a com-
plete conscience. In particular, let us suppose that Lewis is correct
that there are certain universally (or nearly universally) held moral
precepts. Is there any way such universality could be explained other
than as a result of divine activity?
Enter Hume. Hume agrees with Lewis that certain moral beliefs
are widely shared but, like Russell, seeks a naturalistic explanation
for our moral beliefs. He seeks to explain the universality of certain
moral judgments by appealing to certain emotions (or, more pre-
cisely, dispositions to feel certain emotions) that he thinks are part
of human nature. The most important of these is what Hume calls
“benevolence” or “humanity,” which he characterizes as “friendship
for human kind.”72 Humanity is universal in two ways. First, it is uni-
versal in its distribution. The tendency to feel sympathy for others is
common to all human beings. Second, it is universal in its object. We
feel this friendship even toward those who have nothing to do with
us. As Hume puts it, this type of friendship is “so universal and com-
prehensive as to extend to all mankind, and render the actions and
conduct, even of the persons the most remote, an object of applause
Beyond Nature

or censure, according as they agree or disagree with that rule of right
which is established.”73 In Hume™s eyes, moral judgments are rooted
in the emotions, and the emotional disposition he calls “humanity”
explains why certain moral judgements are also universal:

[I]f you represent a tyrannical, insolent, or barbarous behavior, in any
country or in any age of the world, I soon carry my eye to the per-
nicious tendency of such a conduct, and feel the sentiment of repug-
nance and displeasure towards it. No character can be so remote as to
be, in this light, wholly indifferent to me.74

Hume™s tyrant is universally condemned, Hume thinks, because he
invokes the same emotional response from all who re¬‚ect on him.
Because “the humanity of one man is the humanity of every one,”
all who re¬‚ect on the tyrant feel repugnance and displeasure. This
repugnance in turn produces a (negative) moral judgment of the
tyrant™s character. Thus, the universal moral judgment is ultimately
rooted in the humanity common to all human beings.
A second important universal emotional disposition, in Hume™s
eyes, is what he calls “the love of fame.”75 Hume maintains that a
concern to be well regarded by others tends to produce virtue:

By our continual and earnest pursuit of a character, a name, a repu-
tation in the world, we bring our own deportment and conduct fre-
quently in review, and consider how they appear in the eyes of those
who approach and regard us. This constant habit of surveying our-
selves, as it were, in re¬‚ection, keeps alive all the sentiments of right
and wrong, and begets, in noble natures, a certain reverence for them-
selves as well as others, which is the surest guardian of every virtue.76

The passage indicates that the love of fame, working in conjunction
with humanity, plays an important role in producing the moral judg-
ments we make about ourselves. Our love of fame leads us to engage
in frequent self-scrutiny in which we consider ourselves from the
outside, as we appear to others. Once we take this standpoint with
respect to ourselves, our humanity generates moral judgments about
ourselves in much the same way that it generates such judgments
about others.
God and the Reach of Reason

Hume™s remarks are insightful and plausible; but do they make
trouble for Lewis™s moral argument? In the end, I think they do not,
for the following reason. We started out wondering if there might be
an alternative explanation for certain Lewisian moral phenomena
other than activity on the part of a Higher Power. Hume tries to
explain why certain moral judgements and emotions are universal
by appealing to the existence of universal emotional dispositions. But
this just seems to push the question back one step: Why, after all, are
the emotional dispositions Hume identi¬es part of human nature?
What accounts for their universality? Speci¬cally, can the emotional
dispositions universal to human beings be explained other than as a
product of a Higher Power?77
Some contemporary writers have suggested that the relatively
new ¬eld of evolutionary psychology might provide just such an
explanation.78 To understand the basic elements of such an explana-
tion, we must ¬rst understand what is required to provide an evo-
lutionary explanation for the widespread presence of a given trait.
Suppose we observe that all the members of a given species or
population possesses a given trait T. A crucial element of an evolu-
tionary explanation of the widespread presence of T in the population
is support for the claim that, everything else being equal, individual
organisms that possess T are more likely to survive and reproduce
than are individual organisms that lack T.79 The simplest and most
intuitive examples of this sort of explanation involve physical traits.
For instance, imagine a species of birds that subsists on a certain
kind of seed. Suppose that birds of the species in question live in an
environment in which the seeds are found only at the bottom of rel-
atively deep and narrow holes in the ground. Everything else being
equal, birds with longer and narrower beaks will be more likely to
survive and reproduce because they will be better able to reach the
seeds than will birds with shorter or fatter beaks. Birds with shorter
or fatter beaks would be at a disadvantage in the struggle for limited
resources and hence would tend to die out. This sheds light on why
all the birds in the species have long, narrow beaks.
The central idea of evolutionary psychology is that this type of
explanation can be applied not just to physical traits but also to
Beyond Nature

psychological traits. The most interesting (and controversial) appli-
cation of evolutionary psychology is to human beings. It has been
suggested that certain apparently universal human psychological dis-
positions or tendencies can be explained in evolutionary terms. The
contemporary evolutionary psychologist David Buss offers the fol-
lowing illustration:

Consider a common observation that has been documented by scien-
ti¬c research: A woman™s physical appearance is a signi¬cant part of
her desirability to men. . . . The most widely advocated evolutionary
hypothesis is that a woman™s appearance provides a wealth of clues
to her fertility. . . . Over evolutionary time, men who were drawn to
women showing these fertility cues would have outreproduced men
who were drawn to women lacking fertility clues, or who were indif-
ferent to a woman™s physical appearance altogether.80

Some have suggested that dispositions to form certain moral beliefs
or to experience certain moral emotions can be explained in this way,
and evolutionary psychologists have devoted much attention to such
topics. Consider, for example, the disposition to feel gratitude and
warmth toward those who have treated us kindly or fairly and the
disposition to feel outrage and anger toward those who have treated
us unkindly or unfairly. If someone keeps a promise made to us, we
are likely to feel positively about that person and hence to be more
likely to trust her in the future; similarly, we are likely to feel nega-
tively about those who break promises and to be less likely to trust
them. The actions of others produce emotional responses within us,
and these emotional responses in turn in¬‚uence our behavior. For
example, the emotional tendencies I have just described incline us to
engage in “reciprocal altruism” “ cooperation with those who have
proven to be trustworthy partners in the past. More precisely, these
emotional dispositions incline us to follow the “TIT FOR TAT” strat-
egy (named after a computer program written by Anatol Rapaport).
The essence of this strategy is to cooperate with those who have
cooperated with us in the past and to refrain from cooperating with
those who have cheated us in the past.81
The emotional dispositions I have described can seem so natural
and obvious as not to need any explanation; of course we get mad
God and the Reach of Reason

when others cheat us! But the apparent naturalness and obvious-
ness of such reactions supports the claim that they are part of human
nature and therefore good candidates for evolutionary explanation.
Imagine a person who never gets mad when others cheat him. Such a
person will be much easier to take advantage of than someone who
experiences the normal human response to being cheated. Thus,
everything else being equal, a person with the normal emotional
disposition will be more likely to survive and reproduce than the
person incapable of moral outrage. Similarly, a person who remains
indifferent to those who prove to be trustworthy attains fewer of the
bene¬ts of cooperation than a person who experiences the normal
positive feelings of warmth toward those who are trustworthy.
Again, the normal emotional response provides an evolutionary
advantage. Thus, we can see why the normal human emotional dis-
positions would have been selected for by evolution.82
Another moral emotion, closely related to conscience, is guilt. If
those around us tend to follow the TIT FOR TAT strategy, then it
may be in our own best interest to keep our promises and not cheat
others. The emotion of guilt may motivate us to do precisely this; if
we feel guilty (an unpleasant experience) when we cheat, we will be
less likely to do so. And this may make us more likely to attain the
bene¬ts of cooperation. Robert Wright suggests that guilt may also
play a second role:

[G]uilt, which may originally have had the simple role of prompting
payment of overdue debts, could begin to serve a second function:
prompting the preemptive confession of cheating that seems on the
verge of discovery. (Ever notice how guilt does bear a certain correla-
tion with the likelihood of getting caught?)83

Confessing before we are caught will probably be better for us than
merely getting caught; as La Rochefoucauld observes, “[o]ur repen-
tance is less a regret for ills we have caused than a fear of ills we may
encounter.”84 This illustrates another way in which the disposition
to feel guilty in certain situations, while sometimes unpleasant for
the person who has it, can also be advantageous for that person.
Beyond Nature

Thus, we can see why the disposition to feel guilt under the right
circumstances might have been selected for.85
This brief discussion of evolutionary psychology is intended to pro-
vide nothing more than a sketch of how some Lewisian moral phe-
nomena might be explained in a nontheistic fashion. To the extent
that this sort of explanation is plausible, Lewis™s moral argument is
weakened. It is important to note, however, that this kind of expla-
nation does nothing to account for the existence of moral facts. In
Lewis™s terminology, evolutionary explanations may account for our
beliefs and emotions regarding the Law of Nature, but they shed no
light whatsoever on the origin of the Law itself. In general, explaining
why everyone (or almost everyone) believes p is quite different from
explaining why p is true, and this is so when it comes to moral facts.
Explaining why most or all people have certain moral beliefs does
not suf¬ce to explain why the corresponding moral facts are true.
For instance, explaining why most people believe torturing innocent
children just for fun is wrong is one thing; explaining why tortur-
ing innocent children just for fun really is wrong is something else
This distinction is of crucial importance and is often overlooked,
so it is worth pausing a moment to emphasize it. Whenever someone
tells you that she is going to “explain morality,” listen carefully with
an ear toward determining just what it is she is trying to explain.
Is she trying to explain why people have certain moral beliefs and
attitudes? Or is she trying to explain why certain moral claims are
true? Evolutionary psychology may provide a plausible nontheistic
explanation of human moral beliefs and emotions.86 But such an
account by itself does not provide a complete nontheistic explana-
tion of Lewisian moral phenomena. It must be supplemented by a
discussion of the possibility of objective moral truth without a theistic
Can objective moral facts exist in a godless universe? I believe
that they can; moreover, I believe that Lewis™s remarks in “The
Poison of Subjectivism” point us toward a plausible atheistic eth-
ical realism. Recall Lewis™s claim in that essay that “[t]he good is
uncreated; it never could have been otherwise; it has in it no shadow
God and the Reach of Reason

of contingency.”87 I suggested earlier that in this context the good =
the moral law. If this is right, then Lewis™s claim is that the moral
law is uncreated and could not have been otherwise than it is.
Contemporary philosophers tend to distinguish between two
kinds of truths: On the one hand there are contingent truths “ truths
that are true but could have been false. On the other hand there are
necessary truths “ truths that are true and must be true. Necessary
truths are truths that simply could not have been false. In the passage
just quoted, Lewis appears to be claiming that at least some ethical
truths are necessary truths; they have, as Lovell puts it, strong modal
The notion that some ethical truths are necessarily true provides
the foundation for one kind of atheistic ethical realism.89 This is
because necessary truths do not require an explanation of their truth.
Indeed, theists routinely exploit this fact to respond to the question of
God™s origin. A common theistic view is that God did not come from
anywhere; God exists necessarily, and hence His existence requires
no explanation. If some ethical truths are necessarily true, then the
atheist can make a similar claim about them: Their truth does not
need an explanation. Recall that Lewis™s moral argument relies on an
inference to the best explanation; if ethical truths require no expla-
nation, this inference is undercut, and Lewis™s argument loses much
of its bite. I think that arguments like Lewis™s derive much of their
force from the sense that moral obligations must have some source,
that they cannot “just exist.” But if the idea that some moral obliga-
tions are grounded in necessary truths is plausible, then obligations
can indeed “just exist.”
Of course, Lewis claims not only that the moral law exists neces-
sarily but that it is identical to God. My claim here is not that Lewis
himself endorsed the idea that the moral law could exist indepen-
dently of God; rather, what I am suggesting is that Lewis™s claim that
the moral law exists necessarily opens the door to such a proposal.
By my lights, the notion of ethical truths that are necessarily true
and are not identical to God is more plausible than the notion that
such truths are identical to God, because the former view lacks the
obscurity and possible incoherence of the second view.
Beyond Nature

One ¬nal Lewisian moral phenomenon remains unaccounted for:
moral knowledge. Suppose the atheistic alternative to Lewis™s view
sketched so far is plausible; nevertheless, nothing has yet been said
that directly addresses the question of how humans living in a godless
universe could come to know objective ethical truths. This is a com-
plicated topic, but here again we can turn to Lewis himself for some
assistance.90 Recall Lewis™s suggestion that “the primary moral prin-
ciples on which all others depend are rationally perceived. . . . Their
intrinsic reasonableness shines by its own light.”91 The atheist can
appeal to this idea as well: At least some moral truths are self-evident
in that they can be known to be true in a direct way without being
inferred from other things that one knows. For instance, I have no
idea how to prove that torturing the innocent just for fun is wrong,
yet I know it. Once I understand what the claim says, I can simply
see that it is true.92 The idea of “just seeing” certain things to be true
can seem mysterious. However, it is hard to see how we could know
anything at all unless at least some things can be seen to be true even
though we cannot prove that they are true. As we will see, this point
is recognized by Lewis, Hume, and Russell.93
One worry that might arise here is that this view con¬‚icts with the
notion that the human mind is a product of evolutionary processes.
After all, evolutionary processes would tend to select for cognitive
mechanisms that produce moral beliefs that make those who have
them more likely to pass on their genes (¬tness-enhancing cognitive
mechanisms) rather than mechanisms that produce moral beliefs
that are true. So, if our minds are products of evolution, does this
not give us reason to believe that the moral claims that we “just see”
to be true are more likely to be advantageous for us to believe than
actually true?94
This worry can be put to rest, I think, by expanding on the fol-
lowing remark by Peter Singer:

Human beings lack the strength of the gorilla, the sharp teeth of the
lion, the speed of the cheetah. Brain power is our specialty. The brain
is a tool for reasoning, and a capacity to reason helps us to survive,
to feed ourselves, and to safeguard our children. . . . But the ability to
reason is a peculiar ability. Unlike strong arms, sharp teeth or ¬‚ashing

God and the Reach of Reason

legs, it can take us to conclusions that we had no desire to reach. For
reason is like an escalator, leading upwards and out of sight. . . . We
have evolved a capacity to reason because it helps us to survive and
reproduce. But if reason is an escalator, then although the ¬rst part of
the journey may help us to survive and reproduce, we may go further
than we needed to go for this purpose alone.95

Allow me to illustrate the sort of thing I think Singer has in mind
here. Each of the following two cognitive capacities seems likely to be
suf¬ciently ¬tness-enhancing to be the sort of mechanism that would
be selected for by evolution. The ¬rst is the capacity to recognize one-
self as a bearer of certain fundamental rights “ for example, the right
not to be killed for no reason and the right not to be exploited by oth-
ers. Beings that recognize that they have such rights are more likely
to resist treatment that would render them less likely to pass on their
genes to the next generation and hence, everything else being equal,
are more likely in fact to pass on their genes to the next generation
than are beings that fail to recognize that they have such rights.
The second ¬tness-enhancing cognitive capacity is the tendency to
recognize that things that are similar to each other with respect to
their observed properties are likely to be similar with respect to their
unobserved properties (or at least the capacity to reason in accordance
with such a principle, even if one does not consciously recognize the
principle). This capacity is advantageous because the principle it is
centered around is true, and failure to reason in accordance with this
principle can be deadly. Failure to infer that these round, shiny, bright
red berries are likely to be poisonous from the fact that those other
round, shiny, bright red berries produced frothing at the mouth and
then death in one™s companion yesterday may well lead to trouble
for oneself today.
Things get interesting when we notice that these two cognitive
capacities together may lead one to infer that all those beings one
encounters that are similar to oneself with respect to their observ-
able properties have the same fundamental rights as oneself. In this
way, cognitive faculties that in general lead us to beliefs that enhance
our ¬tness may nevertheless in particular cases lead us to conclu-
sions that are not ¬tness-enhancing. Recognizing the rights of beings
Beyond Nature

similar to oneself can put quite a damper on things when it comes to
passing one™s genes on to the next generation; for instance, one can-
not help but realize that taking advantage of that oh-so-exploitable
being similar to oneself would be morally wrong. Thus, the claim
that minds produced by evolution would inevitably form only moral
beliefs that enhance the ¬tness of those who hold them is simply
false; the escalator of reason can lead us to moral beliefs that may
actually make those who hold them less likely to pass on their genes
than they would be if they did not hold such beliefs. And, as in the
case just described, evolution-produced minds may also be “wired”
to zero in on moral truths.
Singer™s escalator analogy also suggests an evolutionary expla-
nation for the internal moral con¬‚ict that pervades human life. Our
minds can reason, and hence can lead us to moral beliefs that con¬‚ict
with our genetic ¬tness. Yet those same minds (for obvious evolu-
tionary reasons) also produce desires that are ¬tness-enhancing. And
thus arises the interminable struggle between doing what is right and
doing what we want to do.96 The human struggle to do what we see
to be right is not necessarily a re¬‚ection of a cosmic struggle between
a Good Power and an Evil Power, or of a disastrous Fall away from
God. It may instead be a consequence of the evolutionary processes
that shaped our minds.
Combining the various ideas discussed in this section affords the
atheist the following response to Lewis™s moral argument: There is
an atheistic explanation for Lewisian moral phenomena that is at
least as good as Lewis™s preferred Higher Power-based explanation.
The atheistic explanation is as follows: Some ethical facts are neces-
sarily true and hence require no explanation; the remaining ethical
facts are contingent and follow from the necessarily true ethical facts
together with certain contingent truths. (For example: Suppose that
it is necessarily true that torturing the innocent just for fun is morally
wrong, and that it is contingently true that by pushing a certain but-
ton, Bob would be torturing the innocent just for fun; from this
follows the (contingent) ethical truth that it would be wrong for Bob
to push the button). Human moral beliefs, knowledge, and emo-
tions are products of the complex, sophisticated human mind, the
God and the Reach of Reason

basic capacities of which can be accounted for in terms of evolution-
ary processes. Even without God, there is real meaning, value, and
morality in the universe, and our imperfect, evolutionarily shaped
minds are able to provide us with at least partial knowledge of the
moral structure of the godless universe. The “escalator of reason”
may even give us knowledge of moral truths that, from the stand-
point of ¬tness, we would be better off not knowing. Hence, there
is no need to posit a Higher Power to account for Lewisian moral
The ultimate conclusion of our discussion of Lewis™s moral argu-
ment is as follows: In order to avoid Russell™s (¬rst) objection, Lewis
must reject this principle:

(RC) The only way a being (even God) can be good is by conform-
ing its actions to a moral law of which it is not the author.

Lewis™s writings suggest three alternatives to (RC):

(LA1) Being identical to the moral law is a way of being good.
(LA2) Loving love, fair play, unsel¬shness, courage, good faith,
honesty, and truthfulness is a way of being good.
(LA3) Desiring that human beings attain genuine happiness (that
they freely love God and strive to become Christlike) is a way
of being good.

Two of these, (LA2) and (LA3), render Lewis™s attack on Dualism
ineffective, and this in turn renders his moral argument vulnerable to
the problem posed by Dualism described in the previous section. On
the other hand, the ideas that underlie (LA1) point toward another
objection to the moral argument, one rooted not in Dualism but
in atheism. The atheistic moral realism sketched in this section, if
plausible, casts serious doubt on the second premise of Lewis™s moral

2. The best explanation of the existence of Lewisian moral phe-
nomena is the existence of a Higher Power that created the uni-
Beyond Nature

My own view is that the contribution of Lewisian moral phenomena
to Lewis™s cumulative case for a Higher Power is weak; I believe that
objective morality is not the thorn in the side of atheism that it is
often thought to be. We have also seen that the phenomenon of
psychopathy does not ¬t particularly well with Lewis™s overall view
of things. While the existence of psychopaths is not decisive evidence
against Lewis™s position, it is one signi¬cant factor that must be taken
into account when weighing the evidence for and against various
worldviews. Of course, there is undoubtedly more to be said here,
and readers must draw their own conclusions. In any case, we now
turn our attention to a second aspect of human nature that Lewis
believes points toward the existence of a Higher Power: our ability
to reason.


In Part IX of Hume™s Dialogues, Philo tentatively suggests the follow-
ing hypothesis:

Is it not probable . . . that the whole economy of the universe is con-
ducted by . . . necessity, though no human algebra can furnish a key
which solves the dif¬culty? And instead of admiring the order of nat-
ural beings, may it not happen that, could we penetrate into the inti-
mate nature of bodies, we should clearly see why it was absolutely
impossible they could ever admit of any other disposition?98

In Miracles, Lewis™s third and ¬nal book-length work in Christian
apologetics, he seeks to refute precisely this hypothesis, which he
labels “Naturalism”:

[B]y Naturalism we mean the doctrine that only Nature “ the whole
interlocked system “ exists. And if that were true, every thing and
event would, if we knew enough, be explicable without remain-
der . . . as a necessary product of the system. The whole system being
what it is, it ought to be a contradiction in terms if you were not
reading this book at the moment.99

The ultimate conclusion of Lewis™s argument against Naturalism is
that there is, in addition to Nature, “an eternal, self-existent, rational
God and the Reach of Reason

Being, whom we call God.”100 The heart of this argument is devel-
oped in the third chapter of Miracles. It is widely known among Lewis
scholars that Lewis wrote two versions of this chapter. Lewis revised
the chapter after a famous encounter with the philosopher G. E. M.
Anscombe on February 2, 1948, at a meeting of the Oxford Socratic
Club.101 On that occasion, Anscombe criticized the ¬rst version of
Chapter 3 of Miracles. Although varying opinions about how dam-
aging Anscombe™s criticisms were to Lewis™s original argument have
been put forth, it is clear that Lewis made signi¬cant revisions to the
chapter after the debate with Anscombe.102 Anscombe herself says of
the revised version of the chapter that “[t]he last ¬ve pages of the old
chapter have been replaced by ten pages of the new . . . the rewritten
version is much less slick and avoids some of the mistakes of the ear-
lier one: it is much more of a serious investigation.”103 I will focus on
the revised version of the chapter, since presumably Lewis took this
to be the strongest presentation of the argument.104 In my view, the
revised chapter is among the most dif¬cult of Lewis™s philosophical
writings to understand, so we will need to do some work to get clear
on exactly how the argument in question is supposed to work.
Before turning to my analysis of Lewis™s argument, I must point
out that there is already an entire book devoted to Lewis™s argument
from reason. This book is Victor Reppert™s C. S. Lewis™s Dangerous Idea.
In that work, Reppert develops six distinct versions of the argument
from reason that he claims are at least suggested by things Lewis says
in various places. I will not undertake a discussion of Reppert™s six
arguments. Instead, I will focus on Lewis™s discussion of the argument
from reason in Miracles, which is his most extended and developed
presentation of the argument. Readers who are interested in other
ways the argument could be developed are encouraged to read Rep-
pert™s book, which is quite well done.105
Let us begin our examination of the argument from reason by
considering a distinction that is crucial to the argument. Lewis dis-
tinguishes two ways that a pair of things can be related to each
other. One relation that can hold between two things is causation
(which Lewis labels “Cause and Effect”). To illustrate this relation,
Lewis uses the sentence “Grandfather is ill today because he ate
Beyond Nature

lobster yesterday.”106 Here, “because” indicates causation; grandfa-
ther™s eating the lobster makes him feel ill. A second relation that can
hold between two things is entailment (which Lewis labels “Ground
and Consequent”). To illustrate this relation, Lewis uses the sentence
“Grandfather must be ill today because he hasn™t gotten up yet (and
we know he is an invariably early riser when he is well).”107 Here,
“because” indicates entailment. The claim (a) whenever grandfather
is not ill he gets up early, and he hasn™t gotten up early today entails
(b) Grandfather is ill today.
With the distinction between causation and entailment in hand,
Lewis argues that if Naturalism is true, “valid” reasoning (by which
Lewis means reasoning that yields knowledge) can occur only if these
two quite different relations can hold between the very same pair of
beliefs.108 To clarify this point and how Lewis arrives at it, it will be
useful to consider an example.
Imagine me, sitting at my desk, struggling to write this exposition
of Lewis™s argument. To give myself a bit of a break, and to verify
that I can still reason properly, imagine that I run through a very
simple argument in my mind. First, I re¬‚ect on (and endorse) the
proposition that (i) all humans are mortal, and I am human. Next, I
re¬‚ect on (and endorse) the proposition that (ii) I am mortal. Let us
refer to my conscious endorsement of proposition (i) as Thought A,
and my conscious endorsement of proposition (ii) as Thought C.
What is required for this simple series of thoughts to constitute
successful reasoning that leads to knowledge? According to Lewis,
one requirement is that the ¬nal thought in the sequence be entailed
by the earlier thought (or, more precisely, that the proposition that
is the object of the ¬nal thought be entailed by the proposition that
is the object of the earlier thought).109 This requirement is met in
our imagined example.
However, Naturalism allegedly imposes an additional requirement
on the series of thoughts under discussion. As Lewis understands
Naturalism, it includes the thesis that every event has a natural cause,
a cause that is itself part of nature. Thus, if Naturalism is true, then
Thought C (which itself must be a part of nature) must have some
cause that is also part of Nature. What could this cause be? The most
God and the Reach of Reason

likely candidate seems to be Thought A. Thus, Naturalism seems to
imply that the series consisting of Thought A and Thought C consti-
tutes reasoning that yields knowledge only if the two very different
relations of causation and entailment both hold between the two
thoughts. As Lewis puts it, “in order for a train of thought to have any
value, these two systems of connection [causation and entailment]
must apply simultaneously to the same series of mental acts.”110 This
gives us the ¬rst premise of Lewis™s argument:

1. If Naturalism is true, then valid reasoning occurs only if one
thought can both entail and cause another thought.

Lewis™s next important move is found in this passage:

We know by experience that a thought does not necessarily cause all,
or even any, of the thoughts which logically stand to it as Consequents
to Ground. We should be in a pretty pickle if we could never think
˜This is glass™ without drawing all the inferences which could be drawn.
It is impossible to draw them all; quite often we draw none. . . . One
thought can cause another not by being, but by being seen to be, a
ground for it.111

Lewis™s point here reveals that the story I just told about conclud-
ing that I am mortal is incomplete. I left out an important part of
the story, namely, the moment when I realized that proposition (i)
(all humans are mortal, and I am human) entails proposition (ii) (I
am mortal). We can label this realization Thought B. If, as Natural-
ism allegedly requires, Thought A causes Thought C, then Thought
B must occur. Lewis™s claim in the ¬nal sentence of the passage I
just quoted implies that it is only through the occurrence of Thought
B that Thought A can cause Thought C. He apparently made much
the same point during the discussion at the meeting of the Oxford
Socratic Club mentioned earlier. On that occasion, Lewis claimed
that “the recognition of a ground could be the cause of assent, and
that assent was only rational when such was its cause.”112 This gives
us a second crucial premise:

2. One thought can both entail and cause another thought only if
the ¬rst thought can be known to entail the second.
Beyond Nature

At this point, Lewis proceeds to argue that Naturalism does not per-
mit the occurrence of thoughts like Thought B. Indeed, Lewis main-
tains that Naturalism does not permit knowledge of any sort. Lewis™s
argument for this claim relies on a certain principle about knowledge:
“An act of knowing must be determined . . . solely by what is known;
we must know it to be thus solely because it is thus.”113 Here, Lewis
endorses a causal principle about knowledge: A person, S, knows a
proposition, p, only if (i) S believes p, and (ii) the complete cause of
S™s belief that p is the truth of p itself. With this principle in hand,
Lewis argues as follows:

If there is nothing but Nature . . . reason must have come into exis-
tence by a historical process. And of course, for the Naturalist, this
process was not designed to produce a mental behaviour that can ¬nd
truth. . . . The type of mental behaviour we now call rational thinking
or inference must therefore have been ˜evolved™ by natural selection,
by the gradual weeding out of types less ¬tted to survive.
Once, then, our thoughts were not rational. . . . Those which had
a cause external to ourselves at all were (like our pains) responses
to stimuli. Now natural selection could operate only by eliminating
responses that were biologically hurtful and multiplying those which
tended to survival. But it is not conceivable that any improvement of
responses could ever turn them into acts of insight, or even remotely
tend to do so. The relation between stimulus and response is utterly
different from that between knowledge and the truth known.114

Strictly speaking, evolutionary theory is not part of Naturalism
as Lewis has de¬ned it, but presumably Lewis™s (reasonable) suppo-
sition is that those who are inclined to accept Naturalism will also
be inclined to accept evolutionary theory. Thus, Lewis supposes that
the Naturalist will likely be committed to the claim that the capacity
for knowledge arose by way of the processes posited by evolutionary
theory, one of which is natural selection. However, Lewis argues,
evolutionary theory has it that the creatures upon which natural
selection initially operated were incapable of knowledge. At best,
they were capable of giving certain responses when certain stimuli
were present. But responding to stimuli is not the same thing as
having knowledge.
God and the Reach of Reason

For instance, imagine a simple creature that feels pain when poked
with a stick. The experience of pain causes the creature to recoil from
the stick. But the creature never forms the belief that it is being poked
by a stick. (Indeed, it forms no belief at all.) There is a mental state
here “ pain “ but no knowledge. The reason there is no knowledge
is that the pain is not about anything. It has no object or content “
it is not what contemporary philosophers of mind call an intentional
state. Instead, it is a subjective feeling that causes a certain action
(recoiling) to occur.
Let us add another mental state to our example. Suppose that the
pain produces a belief that it would be good to recoil, and this belief
in turn produces the act of recoiling. This belief still fails to constitute
knowledge. To see why, recall Lewis™s principle about knowledge: S
knows p only if (i) S believes p, and (ii) the complete cause of S™s
belief that p is the truth of p itself. In the case at hand, the belief
that it would be good to recoil is caused by the truth of the propo-
sition that the creature is being poked by a stick, not by the truth
of the proposition that it would be good to recoil. Hence the second
condition speci¬ed by Lewis™s principle about knowledge is not met.
Again, the point Lewis is trying to make here is that a creature can
have mental states and respond to stimuli in fairly sophisticated ways
without possessing knowledge at all.
This point is essential for Lewis™s argument because an important
part of his argument is the claim that natural selection is incapable
of somehow turning the capacity to respond to certain stimuli into a
capacity for genuine knowledge. If the capacity to respond to stimuli
entailed the capacity for knowledge, then Lewis™s argument would
lose its bite. The naturalist could simply point out that by conceding
that natural selection could produce creatures capable of respond-
ing to stimuli, Lewis would have implicitly conceded that natural
selection could produce creatures capable of knowledge.
In the passage quoted earlier, Lewis claims that the transforma-
tion of (mere) stimuli-responders into genuine knowers by way
of natural selection is “not conceivable.” Therefore, if Naturalism
were true, we might be capable of responding to stimuli, but we

Beyond Nature

would be incapable of genuine knowledge. This yields another pair of

3. If Naturalism is true, then knowledge exists only if natural selec-
tion could produce a capacity for knowledge starting with crea-
tures with no such capacity.
4. But natural selection could not produce a capacity for knowledge
starting with creatures with no such capacity.

Adding these premises to the ¬rst two and making some additional
inferences yields what I take to be Lewis™s main argument in Chap-
ter 3 of Miracles:

1. If Naturalism is true, then valid reasoning occurs only if one
thought can both entail and cause another thought.
2. One thought can both entail and cause another thought only if
the ¬rst thought can be known to entail the second.
3. If Naturalism is true, then knowledge exists only if natural selec-
tion could produce a capacity for knowledge starting with crea-
tures with no such capacity.
4. But natural selection could not produce a capacity for knowledge
starting with creatures with no such capacity.
5. So: If Naturalism is true, then knowledge does not exist (from 3
and 4).
6. If knowledge does not exist, then no thought be known to entail
a second thought.
7. Therefore, if Naturalism is true, then valid reasoning does not
occur (from 1, 2, 5, and 6).

Before considering how the argument progresses from here, it is
worth pausing for a moment to note something odd about the argu-
ment so far. Lewis™s goal in Chapter 3 of Miracles seems to be to show
that Naturalism is self-defeating in the following sense: If it were
true, it could not be known to be true, because there would be no
knowledge at all. Given this, the early stages of the argument (repre-
sented in the formulation presented here by the ¬rst two premises)
are entirely super¬‚uous. This is because premises 3 and 4 by

God and the Reach of Reason

themselves yield the desired conclusion that Naturalism implies that
there is no knowledge.
What is going on here? One hypothesis is that the ¬rst two
premises are what remain of the original argument of Chapter 3
of Miracles, the argument that was criticized by Anscombe at the
meeting of the Oxford Socratic Club in 1948. In order to respond to
Anscombe™s criticism and patch up his argument, Lewis introduced
some additional reasoning, represented by premises 3“5. The new
reasoning, if successful, would indeed patch up the original argu-
ment “ but would also render the original argument entirely unnec-
essary. It may be that Lewis simply failed to notice the latter point.115
In any case, by the end of Chapter 3, Lewis thinks he has estab-
lished that nature alone cannot produce genuine knowledge. Since
knowledge plainly exists, there must be something in addition to
nature that is responsible for knowledge. Since Naturalism entails
that nothing outside of nature exists, Naturalism apparently stands
refuted: “Naturalism . . . offers what professes to be a full account of
our mental behaviour; but this account, on inspection, leaves no
room for the acts of knowing or insight on which the whole value
of our thinking, as a means to truth, depends.”116
In the subsequent chapter of Miracles, Lewis suggests that each
of us possesses a supernatural capacity, reason, which enables us to
have genuine knowledge. Human reason must have some source,
and since it has already been established that human reason cannot
have been produced by nature, it must have a supernatural source.
This supernatural source turns out to be God:


. 3
( 8)