The Evidential Problem of Evil

The evidential problem of evil is the problem of determining whether and, if so,
to what extent the existence of evil (or certain instances, kinds, quantities,
or distributions of evil) constitutes evidence against the existence of God, tha
t is to say, a being perfect in power, knowledge and goodness. Evidential argume
nts from evil attempt to show that, once we put aside any evidence there might b
e in support of the existence of God, it becomes unlikely, if not highly unlikel
y, that the world was created and is governed by an omnipotent, omniscient, and
wholly good being. Such arguments are not to be confused with logical arguments
from evil, which have the more ambitious aim of showing that, in a world in whic
h there is evil, it is logically impossible—and not just unlikely—that God exists.
This entry begins by clarifying some important concepts and distinctions associa
ted with the problem of evil, before providing an outline of one of the more for
ceful and influential evidential arguments developed in contemporary times, name
ly, the evidential argument advanced by William Rowe. Rowe’s argument has occasion
ed a range of responses from theists, including the so-called “skeptical theist” cri
tique (according to which God’s ways are too mysterious for us to comprehend) and
the construction of various theodicies, that is, explanations as to why God perm
its evil. These and other responses to the evidential problem of evil are here s
urveyed and assessed.
Table of Contents
1. Background to the Problem of Evil
1. Orthodox Theism
2. Good and Evil
3. Versions of the Problem of Evil
2. William Rowe’s Evidential Argument from Evil
1. An Outline of Rowe’s Evidential Argument
2. The Theological Premise
3. The Factual Premise
1. Rowe’s Case in Support of the Factual Premise
2. The Inference from P to Q
3. The Skeptical Theist Response
1. Wykstra’s CORNEA Critique
2. Wykstra’s Parent Analogy
3. Alston’s Analogies
4. Building a Theodicy, or Casting Light on the Ways of God
1. What is a Theodicy?
2. Distinguishing a “Theodicy” from a “Defence”
3. Sketch of a Theodicy
5. Further Responses to the Evidential Problem of Evil
6. Conclusion
7. References and Further Reading

1. Background to the Problem of Evil

Before delving into the deep and often murky waters of the problem of evil, it w
ill be helpful to provide some philosophical background to this venerable subjec
t. The first and perhaps most important step of this stage-setting process will
be to identify and clarify the conception of God that is normally presupposed in
contemporary debates (at least within the Anglo-American analytic tradition) on
the problem of evil. The next step will involve providing an outline of some im
portant concepts and distinctions, in particular the age-old distinction between
“good” and “evil,” and the more recent distinction between the logical problem of evil
and the evidential problem of evil.
a. Orthodox Theism
The predominant conception of God within the western world, and hence the kind o
f deity that is normally the subject of debate in discussions on the problem of
evil in most western philosophical circles, is the God of “orthodox theism.” Accordi
ng to orthodox theism, there exists just one God, this God being a person or per
son-like. The operative notion, however, behind this form of theism is that God
is perfect, where to be perfect is to be the greatest being possible or, to borr
ow Anselm’s well-known phrase, the being than which none greater can be conceived.
(Such a conception of God forms the starting-point in what has come to be known
as “perfect being theology”; see Morris 1987, 1991, and Rogers 2000). On this view,
God, as an absolutely perfect being, must possess the following perfections or
great-making qualities:
1. omnipotence: This refers to God’s ability to bring about any state of affa
irs that is logically possible in itself as well as logically consistent with hi
s other essential attributes.
2. omniscience: God is omniscient in that he knows all truths or knows all
that is logically possible to know.
3. perfect goodness: God is the source of moral norms (as in divine command
ethics) or always acts in complete accordance with moral norms.
4. aseity: God has aseity (literally, being from oneself, a se esse) – that i
s to say, he is self-existent or ontologically independent, for he does not depe
nd either for his existence or for his characteristics on anything outside himse
lf.
5. incorporeality: God has no body; he is a non-physical spirit but is capa
ble of affecting physical things.
6. eternity: Traditionally, God is thought to be eternal in an atemporal se
nse—that is, God is timeless or exists outside of time (a view upheld by Augustine
, Boethius, and Aquinas). On an alternative view, God’s eternality is held to be t
emporal in nature, so that God is everlasting or exists in time, having infinite
temporal duration in both of the two temporal directions.
7. omnipresence: God is wholly present in all space and time. This is often
interpreted metaphorically to mean that God can bring about an event immediatel
y at any place and time, and knows what is happening at every place and time in
the same immediate manner.
8. perfectly free: God is absolutely free either in the sense that nothing
outside him can determine him to perform a particular action, or in the sense th
at it is always within his power not to do what he does.
9. alone worthy of worship and unconditional commitment: God, being the gre
atest being possible, is the only being fit to be worshipped and the only being
to whom one may commit one’s life without reservation.
The God of traditional theism is also typically accorded a further attribute, on
e that he is thought to possess only contingently:
1. creator and sustainer of the world: God brought the (physical and non-ph
ysical) world into existence, and also keeps the world and every object within i
t in existence. Thus, no created thing could exist at a given moment unless it w
ere at that moment held in existence by God. Further, no created thing could hav
e the causal powers and liabilities it has at a given moment unless it were at t
hat moment supplied with those powers and liabilities by God.
According to orthodox theism, God was free not to create a world. In other words
, there is at least one possible world in which God creates nothing at all. But
then God is a creator only contingently, not necessarily. (For a more comprehens
ive account of the properties of the God of orthodox theism, see Swinburne 1977,
Quinn & Taliaferro 1997: 223-319, and Hoffman & Rosenkrantz 2002.)
b. Good and Evil
Clarifying the underlying conception of God is but the first step in clarifying
the nature of the problem of evil. To arrive at a more complete understanding of
this vexing problem, it is necessary to unpack further some of its philosophica
l baggage. I turn, therefore, to some important concepts and distinctions associ
ated with the problem of evil, beginning with the ideas of “good” and “evil.”
The terms “good” and “evil” are, if nothing else, notoriously difficult to define. Some
account, however, can be given of these terms as they are employed in discussion
s of the problem of evil. Beginning with the notion of evil, this is normally gi
ven a very wide extension so as to cover everything that is negative and destruc
tive in life. The ambit of evil will therefore include such categories as the ba
d, the unjust, the immoral, and the painful. An analysis of evil in this broad s
ense may proceed as follows:
An event may be categorized as evil if it involves any of the following:
1. some harm (whether it be minor or great) being done to the physical and/
or psychological well-being of a sentient creature;
2. the unjust treatment of some sentient creature;
3. loss of opportunity resulting from premature death;
4. anything that prevents an individual from leading a fulfilling and virtu
ous life;
5. a person doing that which is morally wrong;
6. the “privation of good.”
Condition (a) captures what normally falls under the rubric of pain as a physica
l state (e.g., the sensation you feel when you have a toothache or broken jaw) a
nd suffering as a mental state in which we wish that our situation were otherwis
e (e.g., the experience of anxiety or despair). Condition (b) introduces the not
ion of injustice, so that the prosperity of the wicked, the demise of the virtuo
us, and the denial of voting rights or employment opportunities to women and bla
cks would count as evils. The third condition is intended to cover cases of unti
mely death, that is to say, death not brought about by the ageing process alone.
Death of this kind may result in loss of opportunity either in the sense that o
ne is unable to fulfill one’s potential, dreams or goals, or merely in the sense t
hat one is prevented from living out the full term of their natural life. This i
s partly why we consider it a great evil if an infant were killed after impactin
g with a train at full speed, even if the infant experienced no pain or sufferin
g in the process. Condition (d) classifies as evil anything that inhibits one fr
om leading a life that is both fulfilling and virtuous – poverty and prostitution
would be cases in point. Condition (e) relates evil to immoral choices or acts.
And the final condition expresses the idea, prominent in Augustine and Aquinas,
that evil is not a substance or entity in its own right, but a privatio boni: th
e absence or lack of some good power or quality which a thing by its nature ough
t to possess.
Paralleling the above analysis of evil, the following account of “good” may be offer
ed:
An event may be categorized as good if it involves any of the following:
1. some improvement (whether it be minor or great) in the physical and/or p
sychological well-being of a sentient creature;
2. the just treatment of some sentient creature;
3. anything that advances the degree of fulfillment and virtue in an indivi
dual’s life;
4. a person doing that which is morally right;
5. the optimal functioning of some person or thing, so that it does not lac
k the full measure of being and goodness that ought to belong to it.
Turning to the many varieties of evil, the following have become standard in the
literature:
Moral evil. This is evil that results from the misuse of free will on the part o
f some moral agent in such a way that the agent thereby becomes morally blamewor
thy for the resultant evil. Moral evil therefore includes specific acts of inten
tional wrongdoing such as lying and murdering, as well as defects in character s
uch as dishonesty and greed.
Natural evil. In contrast to moral evil, natural evil is evil that results from
the operation of natural processes, in which case no human being can be held mor
ally accountable for the resultant evil. Classic examples of natural evil are na
tural disasters such as cyclones and earthquakes that result in enormous sufferi
ng and loss of life, illnesses such as leukemia and Alzheimer’s, and disabilities
such as blindness and deafness.
An important qualification, however, must be made at this point. A great deal of
what normally passes as natural evil is brought about by human wrongdoing or ne
gligence. For example, lung cancer may be caused by heavy smoking; the loss of l
ife occasioned by some earthquakes may be largely due to irresponsible city plan
ners locating their creations on faults that will ultimately heave and split; an
d some droughts and floods may have been prevented if not for the careless way w
e have treated our planet. As it is the misuse of free will that has caused thes
e evils or contributed to their occurrence, it seems best to regard them as mora
l evils and not natural evils. In the present work, therefore, a natural evil wi
ll be defined as an evil resulting solely or chiefly from the operation of the l
aws of nature. Alternatively, and perhaps more precisely, an evil will be deemed
a natural evil only if no non-divine agent can be held morally responsible for
its occurrence. Thus, a flood caused by human pollution of the environment will
be categorized a natural evil as long as the agents involved could not be held m
orally responsible for the resultant evil, which would be the case if, for insta
nce, they could not reasonably be expected to have foreseen the consequences of
their behavior.
A further category of evil that has recently played an important role in discuss
ions on the problem of evil is horrendous evil. This may be defined, following M
arilyn Adams (1999: 26), as evil “the participation in which (that is, the doing o
r suffering of which) constitutes prima facie reason to doubt whether the partic
ipant’s life could (given their inclusion in it) be a great good to him/her on the
whole.” As examples of such evil, Adams lists “the rape of a woman and axing off of
her arms, psycho-physical torture whose ultimate goal is the disintegration of
personality, betrayal of one’s deepest loyalties, child abuse of the sort describe
d by Ivan Karamazov, child pornography, parental incest, slow death by starvatio
n, the explosion of nuclear bombs over populated areas” (p.26).
A horrendous evil, it may be noted, may be either a moral evil (e.g., the Holoca
ust of 1939-45) or a natural evil (e.g., the Lisbon earthquake of 1755). It is a
lso important to note that it is the notion of a “horrendous moral evil” that compor
ts with the current, everyday use of “evil” by English speakers. When we ordinarily
employ the word “evil” today we do not intend to pick out something that is merely b
ad or very wrong (e.g., a burglary), nor do we intend to refer to the death and
destruction brought about by purely natural processes (we do not, for example, t
hink of the 2004 Asian tsunami disaster as something that was “evil”). Instead, the
word “evil” is reserved in common usage for events and people that have an especiall
y horrific moral quality or character.
Clearly, the problem of evil is at its most difficult when stated in terms of ho
rrendous evil (whether of the moral or natural variety), and as will be seen in
Section II below, this is precisely how William Rowe’s statement of the evidential
problem of evil is formulated.
Finally, these notions of good and evil indicate that the problem of evil is int
imately tied to ethics. One’s underlying ethical theory may have a bearing on one’s
approach to the problem of evil in at least two ways.
Firstly, one who accepts either a divine command theory of ethics or non-realism
in ethics is in no position to raise the problem of evil, that is, to offer the
existence of evil as at least a prima facie good reason for rejecting theism. T
his is because a divine command theory, in taking morality to be dependent upon
the will of God, already assumes the truth of that which is in dispute, namely,
the existence of God (see Brown 1967). On the other hand, non-realist ethical th
eories, such as moral subjectivism and error-theories of ethics, hold that there
are no objectively true moral judgments. But then a non-theist who also happens
to be a non-realist in ethics cannot help herself to some of the central premis
es found in evidential arguments from evil (such as “If there were a perfectly goo
d God, he would want a world with no horrific evil in it”), as these purport to be
objectively true moral judgments (see Nelson 1991). This is not to say, however
, that atheologians such as David Hume, Bertrand Russell and J.L. Mackie, each o
f whom supported non-realism in ethics, were contradicting their own meta-ethics
when raising arguments from evil – at least if their aim was only to show up a co
ntradiction in the theist’s set of beliefs.
Secondly, the particular normative ethical theory one adopts (e.g., consequentia
lism, deontology, virtue ethics) may influence the way in which one formulates o
r responds to an argument from evil. Indeed, some have gone so far as to claim t
hat evidential arguments from evil usually presuppose the truth of consequential
ism (see, for example, Reitan 2000). Even if this is not so, it seems that the a
doption of a particular theory in normative ethics may render the problem of evi
l easier or harder, or at least delimit the range of solutions available. (For a
n excellent account of the difficulties faced by theists in relation to the prob
lem of evil when the ethical framework is restricted to deontology, see McNaught
on 1994.)
c. Versions of the Problem of Evil
The problem of evil may be described as the problem of reconciling belief in God
with the existence of evil. But the problem of evil, like evil itself, has many
faces. It may, for example, be expressed either as an experiential problem or a
s a theoretical problem. In the former case, the problem is the difficulty of ad
opting or maintaining an attitude of love and trust toward God when confronted b
y evil that is deeply perplexing and disturbing. Alvin Plantinga (1977: 63-64) p
rovides an eloquent account of this problem:
The theist may find a religious problem in evil; in the presence of his own suff
ering or that of someone near to him he may find it difficult to maintain what h
e takes to be the proper attitude towards God. Faced with great personal sufferi
ng or misfortune, he may be tempted to rebel against God, to shake his fist in G
od’s face, or even to give up belief in God altogether… Such a problem calls, not fo
r philosophical enlightenment, but for pastoral care. (emphasis in the original)
By contrast, the theoretical problem of evil is the purely “intellectual” matter of
determining what impact, if any, the existence of evil has on the truth-value or
the epistemic status of theistic belief. To be sure, these two problems are int
erconnected – theoretical considerations, for example, may color one’s actual experi
ence of evil, as happens when suffering that is better comprehended becomes easi
er to bear. In this article, however, the focus will be exclusively on the theor
etical dimension. This aspect of the problem of evil comes in two broad varietie
s: the logical problem and the evidential problem.
The logical version of the problem of evil (also known as the a priori version a
nd the deductive version) is the problem of removing an alleged logical inconsis
tency between certain claims about God and certain claims about evil. J.L. Macki
e (1955: 200) provides a succinct statement of this problem:
In its simplest form the problem is this: God is omnipotent; God is wholly good;
and yet evil exists. There seems to be some contradiction between these three p
ropositions, so that if any two of them were true the third would be false. But
at the same time all three are essential parts of most theological positions: th
e theologian, it seems, at once must adhere and cannot consistently adhere to al
l three. (emphases in the original)
In a similar vein, H.J. McCloskey (1960: 97) frames the problem of evil as follo
ws:
Evil is a problem for the theist in that a contradiction is involved in the fact
of evil, on the one hand, and the belief in the omnipotence and perfection of G
od on the other. (emphasis mine)
Atheologians like Mackie and McCloskey, in maintaining that the logical problem
of evil provides conclusive evidence against theism, are claiming that theists a
re committed to an internally inconsistent set of beliefs and hence that theism
is necessarily false. More precisely, it is claimed that theists commonly accept
the following propositions:
1. God exists
2. God is omnipotent
3. God is omniscient
4. God is perfectly good
5. Evil exists.
Propositions (11)-(14) form an essential part of the orthodox conception of God,
as this has been explicated in Section 1 above. But theists typically believe t
hat the world contains evil. The charge, then, is that this commitment to (15) i
s somehow incompatible with the theist’s commitment to (11)-(14). Of course, (15)
can be specified in a number of ways – for example, (15) may refer to the existenc
e of any evil at all, or a certain amount of evil, or particular kinds of evil,
or some perplexing distributions of evil. In each case, a different version of t
he logical problem of evil, and hence a distinct charge of logical incompatibili
ty, will be generated.
The alleged incompatibility, however, is not obvious or explicit. Rather, the cl
aim is that propositions (11)-(15) are implicitly contradictory, where a set S o
f propositions is implicitly contradictory if there is a necessary proposition p
such that the conjunction of p with S constitutes a formally contradictory set.
Those who advance logical arguments from evil must therefore add one or more ne
cessary truths to the above set of five propositions in order to generate the fa
tal contradiction. By way of illustration, consider the following additional pro
positions that may be offered:
1. A perfectly good being would want to prevent all evils.
2. An omniscient being knows every way in which evils can come into existen
ce.
3. An omnipotent being who knows every way in which an evil can come into e
xistence has the power to prevent that evil from coming into existence.
4. A being who knows every way in which an evil can come into existence, wh
o is able to prevent that evil from coming into existence, and who wants to do s
o, would prevent the existence of that evil.
From this set of auxiliary propositions, it clearly follows that
1. If there exists an omnipotent, omniscient, and perfectly good being, the
n no evil exists.
It is not difficult to see how the addition of (16)-(20) to (11)-(15) will yield
an explicit contradiction, namely,
1. Evil exists and evil does not exist.
If such an argument is sound, theism will not so much lack evidential support, b
ut would rather be, as Mackie (1955: 200) puts it, “positively irrational.” For more
discussion, see the article The Logical Problem of Evil.
The subject of this article, however, is the evidential version of the problem o
f evil (also called the a posteriori version and the inductive version), which s
eeks to show that the existence evil, although logically consistent with the exi
stence of God, counts against the truth of theism. As with the logical problem,
evidential formulations may be based on the sheer existence of evil, or certain
instances, types, amounts, or distributions of evil. Evidential arguments from e
vil may also be classified according to whether they employ (i) a direct inducti
ve approach, which aims at showing that evil counts against theism, but without
comparing theism to some alternative hypothesis; or (ii) an indirect inductive a
pproach, which attempts to show that some significant set of facts about evil co
unts against theism, and it does this by identifying an alternative hypothesis t
hat explains these facts far more adequately than the theistic hypothesis. The f
ormer strategy, as will be seen in Section II, is employed by William Rowe, whil
e the latter strategy is exemplified best in Paul Draper’s 1989 paper, “Pain and Ple
asure: An Evidential Problem for Theists”. (A useful taxonomy of evidential argume
nts from evil can be found in Russell 1996: 194 and Peterson 1998: 23-27, 69-72.
)
Evidential arguments purport to show that evil counts against theism in the sens
e that the existence of evil lowers the probability that God exists. The strateg
y here is to begin by putting aside any positive evidence we might think there i
s in support of theism (e.g., the fine-tuning argument) as well as any negative
evidence we might think there is against theism (that is, any negative evidence
other than the evidence of evil). We therefore begin with a “level playing field” by
setting the probability of God’s existing at 0.5 and the probability of God’s not e
xisting at 0.5 (compare Rowe 1996: 265-66; it is worth noting, however, that thi
s “level playing field” assumption is not entirely uncontroversial: see, for example
, the objections raised by Jordan 2001 and Otte 2002: 167-68). The aim is to the
n determine what happens to the probability value of “God exists” once we consider t
he evidence generated by our observations of the various evils in our world. The
central question, therefore, is: Grounds for belief in God aside, does evil ren
der the truth of atheism more likely than the truth of theism? (A recent debate
on the evidential problem of evil was couched in such terms: see Rowe 2001a: 124
-25.) Proponents of evidential arguments are therefore not claiming that, even i
f we take into account any positive reasons there are in support of theism, the
evidence of evil still manages to lower the probability of God’s existence. They a
re only making the weaker claim that, if we temporarily set aside such positive
reasons, then it can be shown that the evils that occur in our world push the pr
obability of God’s existence significantly downward.
But if evil counts against theism by driving down the probability value of “God ex
ists” then evil constitutes evidence against the existence of God. Evidential argu
ments, therefore, claim that there are certain facts about evil that cannot be a
dequately explained on a theistic account of the world. Theism is thus treated a
s a large-scale hypothesis or explanatory theory which aims to make sense of som
e pertinent facts, and to the extent that it fails to do so it is disconfirmed.
In evidential arguments, however, the evidence only probabilifies its conclusion
, rather than conclusively verifying it. The probabilistic nature of such argume
nts manifests itself in the form of a premise to the effect that “It is probably t
he case that some instance (or type, or amount, or pattern) of evil E is gratuit
ous.” This probability judgment usually rests on the claim that, even after carefu
l reflection, we can see no good reason for God’s permission of E. The inference f
rom this claim to the judgment that there exists gratuitous evil is inductive in
nature, and it is this inductive step that sets the evidential argument apart f
rom the logical argument.
2. William Rowe’s Evidential Argument from Evil
Evidential arguments from evil seek to show that the presence of evil in the wor
ld inductively supports or makes likely the claim that God (or, more precisely,
the God of orthodox theism) does not exist. A variety of evidential arguments ha
ve been formulated in recent years, but here I will concentrate on one very infl
uential formulation, namely, that provided by William Rowe. Rowe’s version of the
evidential argument has received much attention since its formal inception in 19
78, for it is often considered to be the most cogent presentation of the evident
ial problem of evil. James Sennett (1993: 220), for example, views Rowe’s argument
as “the clearest, most easily understood, and most intuitively appealing of those
available.” Terry Christlieb (1992: 47), likewise, thinks of Rowe’s argument as “the
strongest sort of evidential argument, the sort that has the best chance of succ
ess.” It is important to note, however, that Rowe’s thinking on the evidential probl
em of evil has developed in significant ways since his earliest writings on the
subject, and two (if not three) distinct evidential arguments can be identified
in his work. Here I will only discuss that version of Rowe’s argument that receive
d its first full-length formulation in Rowe (1978) and, most famously, in Rowe (
1979), and was successively refined in the light of criticisms in Rowe (1986), (
1988), (1991), and (1995), before being abandoned in favour of a quite different
evidential argument in Rowe (1996).
a. An Outline of Rowe’s Argument
In presenting his evidential argument from evil in his justly celebrated 1979 pa
per, “The Problem of Evil and Some Varieties of Atheism”, Rowe thinks it best to foc
us on a particular kind of evil that is found in our world in abundance. He ther
efore selects “intense human and animal suffering” as this occurs on a daily basis,
is in great plenitude in our world, and is a clear case of evil. More precisely,
it is a case of intrinsic evil: it is bad in and of itself, even though it some
times is part of, or leads to, some good state of affairs (Rowe 1979: 335). Rowe
then proceeds to state his argument for atheism as follows:
1. There exist instances of intense suffering which an omnipotent, omniscie
nt being could have prevented without thereby losing some greater good or permit
ting some evil equally bad or worse.
2. An omniscient, wholly good being would prevent the occurrence of any int
ense suffering it could, unless it could not do so without thereby losing some g
reater good or permitting some evil equally bad or worse.
3. (Therefore) There does not exist an omnipotent, omniscient, wholly good
being. (Rowe 1979: 336)
This argument, as Rowe points out, is clearly valid, and so if there are rationa
l grounds for accepting its premises, to that extent there are rational grounds
for accepting the conclusion, that is to say, atheism.
b. The Theological Premise
The second premise is sometimes called “the theological premise” as it expresses a b
elief about what God as a perfectly good being would do under certain circumstan
ces. In particular, this premise states that if such a being knew of some intens
e suffering that was about to take place and was in a position to prevent its oc
currence, then it would prevent it unless it could not do so without thereby los
ing some greater good or permitting some evil equally bad or worse. Put otherwis
e, an omnipotent, omniscient, wholly good God would not permit any gratuitous ev
il, evil that is (roughly speaking) avoidable, pointless, or unnecessary with re
spect to the fulfillment of God’s purposes.
Rowe takes the theological premise to be the least controversial aspect of his a
rgument. And the consensus seems to be that Rowe is right – the theological premis
e, or a version thereof that is immune from some minor infelicities in the origi
nal formulation, is usually thought to be indisputable, self-evident, necessaril
y true, or something of that ilk. The intuition here, as the Howard-Snyders (199
9: 115) explain, is that “on the face of it, the idea that God may well permit gra
tuitous evil is absurd. After all, if God can get what He wants without permitti
ng some particular horror (or anything comparably bad), why on earth would He pe
rmit it?”
An increasing number of theists, however, are beginning to question Rowe’s theolog
ical premise. This way of responding to the evidential problem of evil has been
described by Rowe as “radical, if not revolutionary” (1991: 79), but it is viewed by
many theists as the only way to remain faithful to the common human experience
of evil, according to which utterly gratuitous evil not only exists but is abund
ant. In particular, some members of the currently popular movement known as open
theism have rallied behind the idea that the theistic worldview is not only com
patible with, but requires or demands, the possibility that there is gratuitous
evil (for the movement’s “manifesto,” see Pinnock et al. 1994; see also Sanders 1998,
Boyd 2000, and Hasker 2004).
Although open theists accept the orthodox conception of God, as delineated in Se
ction I.1 above, they offer a distinct account of some of the properties that ar
e constitutive of the orthodox God. Most importantly, open theists interpret God’s
omniscience in such a way that it does not include either foreknowledge (or, mo
re specifically, knowledge of what free agents other than God will do) or middle
knowledge (that is, knowledge of what every possible free creature would freely
choose to do in any possible situation in which that creature might find itself
). This view is usually contrasted with two other forms of orthodox theism: Moli
nism (named after the sixteenth-century Jesuit theologian Luis de Molina, who de
veloped the theory of middle knowledge), according to which divine omniscience e
ncompasses both foreknowledge and middle knowledge; and Calvinism or theological
determinism, according to which God determines or predestines all that happens,
thus leaving us with either no morally relevant free will at all (hard determin
ism) or free will of the compatibilist sort only (soft determinism).
It is often thought that the Molinist and Calvinist grant God greater providenti
al control over the world than does the open theist. For according to the latter
but not the former, the future is to some degree open-ended in that not even Go
d can know exactly how it will turn out, given that he has created a world in wh
ich there are agents with libertarian free will and, perhaps, indeterminate natu
ral processes. God therefore runs the risk that his creation will come to be inf
ested with gratuitous evils, that is to say, evils he has not intended, decreed,
planned for, or even permitted for the sake of some greater good. Open theists,
however, argue that this risk is kept in check by God’s adoption of various gener
al strategies by which he governs the world. God may, for example, set out to cr
eate a world in which there are creatures who have the opportunity to freely cho
ose their destiny, but he would then ensure that adequate recompense is offered
(perhaps in an afterlife) to those whose lives are ruined (through no fault of t
heir own) by the misuse of others’ freedom (e.g., a child that is raped and murder
ed). Nevertheless, in creating creatures with (libertarian) free will and by inf
using the natural order with a degree of indeterminacy, God relinquishes exhaust
ive knowledge and complete control of all history. The open theist therefore enc
ourages the rejection of what has been called “meticulous providence” (Peterson 1982
: chs 4 & 5) or “the blueprint worldview” (Boyd 2003: ch.2), the view that the world
was created according to a detailed divine blueprint which assigns a specific d
ivine reason for every occurrence in history. In place of this view, the open th
eist presents us with a God who is a risk-taker, a God who gives up meticulous c
ontrol of everything that happens, thus opening himself up to the genuine possib
ility of failure and disappointment – that is to say, to the possibility of gratui
tous evil.
Open theism has sparked much heated debate and has been attacked from many quart
ers. Considered, however, as a response to Rowe’s theological premise, open theism’s
prospects seem dim. The problem here, as critics have frequently pointed out, i
s that the open view of God tends to diminish one’s confidence in God’s ability to e
nsure that his purposes for an individual’s life, or for world history, will be ac
complished (see, e.g., Ware 2000, Ascol 2001: 176-80). The worry is that if, as
open theists claim, God does not exercise sovereign control over the world and t
he direction of human history is open-ended, then it seems that the world is lef
t to the mercy of Tyche or Fortuna, and we are therefore left with no assurance
that God’s plan for the world and for us will succeed. Consider, for example, Eleo
nore Stump’s rhetorical questions, put in response to the idea of a “God of chance” ad
vocated in van Inwagen (1988): “Could one trust such a God with one’s child, one’s lif
e? Could one say, as the Psalmist does, “I will both lay me down in peace and slee
p, for thou, Lord, only makest me dwell in safety’?” (1997: 466, quoting from Psalm
4:8). The answer may in large part depend on the degree to which the world is th
ought to be imbued with indeterminacy or chance.
If, for example, the open theist view introduces a high level of chance into God’s
creation, this would raise the suspicion that the open view reflects an excessi
vely deistic conception of God’s relation to the world. Deism is popularly thought
of as the view that a supreme being created the world but then, like an absente
e landlord, left it to run on its own accord. Deists, therefore, are often accus
ed of postulating a remote and indifferent God, one who does not exercise provid
ential care over his creation. Such a deity, it might be objected, resembles the
open theist’s God of chance. The objection, in other words, is that open theists
postulate a dark and risky universe subject to the forces of blind chance, and t
hat it is difficult to imagine a personal God—that is, a God who seeks to be perso
nally related to us and hence wants us to develop attitudes of love and trust to
wards him – providing us with such a habitat. To paraphrase Einstein, God does not
play dice with our lives.
This, however, need not mean that God does not play dice at all. It is not impos
sible, in other words, to accommodate chance within a theistic world-view. To se
e this, consider a particular instance of moral evil: the rape and murder of a l
ittle girl. It seems plausible that no explanation is available as to why God wo
uld permit this specific evil (or, more precisely, why God would permit this gir
l to suffer then and there and in that way), since any such explanation that is
offered will inevitably recapitulate the explanation offered for at least one of
the major evil-kinds that subsumes the particular evil in question (e.g., the c
lass of moral evils). It is therefore unreasonable to request a reason (even a p
ossible reason) for God’s permission of a particular event that is specific to thi
s event and that goes beyond some general policy or plan God might have for perm
itting events of that kind. If this correct, then there is room for theists to a
ccept the view that at least some evils are chancy or gratuitous in the sense th
at there is no specific reason as to why these evils are permitted by God. Howev
er, this kind of commitment to gratuitous evil is entirely innocuous for propone
nts of Rowe’s theological premise. For one can simply modify this premise so that
it ranges either over particular instances of evil or (to accommodate cases wher
e particular evils admit of no divine justification) over broadly defined evils
or evil-kinds under which the relevant particular evils can be subsumed. And so
a world created by God may be replete with gratuitous evil, as open theists imag
ine, but that need not present a problem for Rowe.
(For a different line of argument in support of the compatibility of theism and
gratuitous evil, see Hasker (2004: chs 4 & 5), who argues that the consequences
for morality would be disastrous if we took Rowe’s theological premise to be true.
For criticisms of this view, see Rowe (1991: 79-86), Chrzan (1994), O’Connor (199
8: 53-70), and Daniel and Frances Howard-Snyder (1999: 119-27).)
c. The Factual Premise
Criticisms of Rowe’s argument tend to focus on its first premise, sometimes dubbed
“the factual premise,” as it purports to state a fact about the world. Briefly put,
the fact in question is that there exist instances of intense suffering which a
re gratuitous or pointless. As indicated above, an instance of suffering is grat
uitous, according to Rowe, if an omnipotent, omniscient being could have prevent
ed it without thereby losing some greater good or permitting some evil equally b
ad or worse. A gratuitous evil, in this sense, is a state of affairs that is not
(logically) necessary to the attainment of a greater good or to the prevention
of an evil at least as bad.
i Rowe’s Case in Support of the Factual Premise
Rowe builds his case in support of the factual premise by appealing to particula
r instances of human and animal suffering, such as the following:
E1: the case of Bambi
“In some distant forest lightning strikes a dead tree, resulting in a forest fire.
In the fire a fawn is trapped, horribly burned, and lies in terrible agony for
several days before death relieves its suffering” (Rowe 1979: 337).
Although this is presented as a hypothetical event, Rowe takes it to be “a familia
r sort of tragedy, played not infrequently on the stage of nature” (1988: 119).
E2: the case of Sue
This is an actual event in which a five-year-old girl in Flint, Michigan was sev
erely beaten, raped and then strangled to death early on New Year’s Day in 1986. T
he case was introduced by Bruce Russell (1989: 123), whose account of it, drawn
from a report in the Detroit Free Press of January 3 1986, runs as follows:
The girl’s mother was living with her boyfriend, another man who was unemployed, h
er two children, and her 9-month old infant fathered by the boyfriend. On New Ye
ar’s Eve all three adults were drinking at a bar near the woman’s home. The boyfrien
d had been taking drugs and drinking heavily. He was asked to leave the bar at 8
:00 p.m. After several reappearances he finally stayed away for good at about 9:
30 p.m. The woman and the unemployed man remained at the bar until 2:00 a.m. at
which time the woman went home and the man to a party at a neighbor’s home. Perhap
s out of jealousy, the boyfriend attacked the woman when she walked into the hou
se. Her brother was there and broke up the fight by hitting the boyfriend who wa
s passed out and slumped over a table when the brother left. Later the boyfriend
attacked the woman again, and this time she knocked him unconscious. After chec
king the children, she went to bed. Later the woman’s 5-year old girl went downsta
irs to go to the bathroom. The unemployed man returned from the party at 3:45 a.
m. and found the 5-year old dead. She had been raped, severely beaten over most
of her body and strangled to death by the boyfriend.
Following Rowe (1988: 120), the case of the fawn will be referred to as “E1″, and th
e case of the little girl as “E2″. Further, following William Alston’s (1991: 32) prac
tice, the fawn will be named “Bambi” and the little girl “Sue”.
Rowe (1996: 264) states that, in choosing to focus on E1 and E2, he is “trying to
pose a serious difficulty for the theist by picking a difficult case of natural
evil, E1 (Bambi), and a difficult case of moral evil, E2 (Sue).” Rowe, then, is at
tempting to state the evidential argument in the strongest possible terms. As on
e commentator has put it, “if these cases of evil [E1 and E2] are not evidence aga
inst theism, then none are” (Christlieb 1992: 47). However, Rowe’s almost exclusive
preoccupation with these two instances of suffering must be placed within the co
ntext of his belief (as expressed in, e.g., 1979: 337-38) that even if we discov
ered that God could not have eliminated E1 and E2 without thereby losing some gr
eater good or permitting some evil equally bad or worse, it would still be unrea
sonable to believe this of all cases of horrendous evil occurring daily in our w
orld. E1 and E2 are thus best viewed as representative of a particular class of
evil which poses a specific problem for theistic belief. This problem is express
ed by Rowe in the following way:
(P) No good state of affairs we know of is such that an omnipotent, omniscient b
eing’s obtaining it would morally justify that being’s permitting E1 or E2. Therefor
e,
(Q) It is likely that no good state of affairs is such that an omnipotent, omnis
cient being’s obtaining it would morally justify that being in permitting E1 or E2
.
P states that no good we know of justifies God in permitting E1 and E2. From thi
s it is inferred that Q is likely to be true, or that probably there are no good
s which justify God in permitting E1 and E2. Q, of course, corresponds to the fa
ctual premise of Rowe’s argument. Thus, Rowe attempts to establish the truth of th
e factual premise by appealing to P.
ii. The Inference from P to Q
At least one question to be addressed when considering this inference is: What e
xactly do P and Q assert? Beginning with P, the central notion here is “a good sta
te of affairs we know of.” But what is it to know of a good state of affairs? Acco
rding to Rowe (1988: 123), to know of a good state of affairs is to (a) conceive
of that state of affairs, and (b) recognize that it is intrinsically good (exam
ples of states that are intrinsically good include pleasure, happiness, love, an
d the exercise of virtue). Rowe (1996: 264) therefore instructs us to not limit
the set of goods we know of to goods that we know have occurred in the past or t
o goods that we know will occur in the future. The set of goods we know of must
also include goods that we have some grasp of, even if we do not know whether th
ey have occurred or ever will occur. For example, such a good, in the case of Su
e, may consist of the experience of eternal bliss in the hereafter. Even though
we lack a clear grasp of what this good involves, and even though we cannot be s
ure that such a good will ever obtain, we do well to include this good amongst t
he goods we know of. A good that we know of, however, cannot justify God in perm
itting E1 or E2 unless that good is actualized at some time.
On what grounds does Rowe think that P is true? Rowe (1988: 120) states that “we h
ave good reason to believe that no good state of affairs we know of would justif
y an omnipotent, omniscient being in permitting either E1 or E2” (emphasis his). T
he good reason in question consists of the fact that the good states of affairs
we know of, when reflecting on them, meet one or both of the following condition
s: either an omnipotent being could obtain them without having to permit E1 or E
2, or obtaining them would not morally justify that being in permitting E1 or E2
(Rowe 1988: 121, 123; 1991: 72).
This brings us, finally, to Rowe’s inference from P to Q. This is, of course, an i
nductive inference. Rowe does not claim to know or be able to prove that cases o
f intense suffering such as the fawn’s are indeed pointless. For as he acknowledge
s, it is quite possible that there is some familiar good outweighing the fawn’s su
ffering and which is connected to that suffering in a way unbeknown to us. Or th
ere may be goods we are not aware of, to which the fawn’s suffering is intimately
connected. But although we do not know or cannot establish the truth of Q, we do
possess rational grounds for accepting Q, and these grounds consist of the cons
iderations adumbrated in P. Thus, the truth of P is taken to provide strong evid
ence for the truth of Q (Rowe 1979: 337).
3. The Skeptical Theist Response
Theism, particularly as expressed within the Judeo-Christian and Islamic religio
ns, has always emphasized the inscrutability of the ways of God. In Romans 11:33
-34, for example, the apostle Paul exclaims: “Oh, the depth of the riches of the w
isdom and knowledge of God! How unsearchable his judgments, and his paths beyond
tracing out! Who has known the mind of the Lord?” (NIV). This emphasis on mystery
and the epistemic distance between God and human persons is a characteristic te
net of traditional forms of theism. It is in the context of this tradition that
Stephen Wykstra developed his well-known CORNEA critique of Rowe’s evidential argu
ment. The heart of Wykstra’s critique is that, given our cognitive limitations, we
are in no position to judge as improbable the statement that there are goods be
yond our ken secured by God’s permission of many of the evils we find in the world
. This position – sometimes labelled “skeptical theism” or “defensive skepticism” – has gen
rated a great deal of discussion, leading some to conclude that “the inductive arg
ument from evil is in no better shape than its late lamented deductive cousin” (Al
ston 1991: 61). In this Section, I will review the challenge posed by this theis
tic form of skepticism, beginning with the critique advanced by Wykstra.
a. Wykstra’s CORNEA Critique
In an influential paper entitled, “The Humean Obstacle to Evidential Arguments fro
m Evil,” Stephen Wykstra raised a formidable objection to Rowe’s inference from P to
Q. Wykstra’s first step was to draw attention to the following epistemic principl
e, which he dubbed “CORNEA” (short for “Condition Of ReasoNable Epistemic Access”):
(C) On the basis of cognized situation s, human H is entitled to claim “It appears
that p” only if it is reasonable for H to believe that, given her cognitive facul
ties and the use she has made of them, if p were not the case, s would likely be
different than it is in some way discernible by her. (Wykstra 1984: 85)
The point behind CORNEA may be easier to grasp if (C) is simplified along the fo
llowing lines:
(C*) H is entitled to infer “There is no x“ from “So far as I can tell, there is no x” o
nly if:
It is reasonable for H to believe that if there were an x, it is likely that she
would perceive (or find, grasp, comprehend, conceive) it.
Adopting terminology introduced by Wykstra (1996), the inference from “So far as I
can tell, there is no x” to “There is no x” may be called a “noseeum inference”: we no se
e ’um, so they ain’t there! Further, the italicized portion in (C*) may be called “the
noseeum assumption,” as anyone who employs a noseeum inference and is justified i
n doing so would be committed to this assumption.
C*, or at least something quite like it, appears unobjectionable. If, for instan
ce, I am looking through the window of my twentieth-floor office to the garden b
elow and I fail to see any caterpillars on the flowers, that would hardly entitl
e me to infer that there are in fact no caterpillars there. Likewise, if a begin
ner were watching Kasparov play Deep Blue, it would be unreasonable for her to i
nfer “I can’t see any way for Deep Blue to get out of check; so, there is none.” Both
inferences are illegitimate for the same reason: the person making the inference
does not have what it takes to discern the sorts of things in question. It is t
his point that C* intends to capture by insisting that a noseeum inference is pe
rmissible only if it is likely that one would detect or discern the item in ques
tion if it existed.
But how does the foregoing relate to Rowe’s evidential argument? Notice, to begin
with, that Rowe’s inference from P to Q is a noseeum inference. Rowe claims in P t
hat, so far as we can see, no goods justify God’s permission of E1 and E2, and fro
m this he infers that no goods whatever justify God’s permission of these evils. A
ccording to Wykstra, however, Rowe is entitled to make this noseeum inference on
ly if he is entitled to make the following noseeum assumption:
If there are goods justifying God’s permission of horrendous evil, it is likely th
at we would discern or be cognizant of such goods.
Call this Rowe’s Noseeum Assumption, or RNA for short. The key issue, then, is whe
ther we should accept RNA. Many theists, led by Stephen Wykstra, have claimed th
at RNA is false (or that we ought to suspend judgement about its truth). They ar
gue that the great gulf between our limited cognitive abilities and the infinite
wisdom of God prevents us (at least in many cases) from discerning God’s reasons
for permitting evil. On this view, even if there are goods secured by God’s permis
sion of evil, it is likely that these goods would be beyond our ken. Alvin Plant
inga (1974: 10) sums up this position well with his rhetorical question: “Why supp
ose that if God does have a reason for permitting evil, the theist would be the
first to know?” (emphasis his). Since theists such as Wykstra and Plantinga challe
nge Rowe’s argument (and evidential arguments in general) by focusing on the limit
s of human knowledge, they have become known as skeptical theists.
I will now turn to some considerations that have been offered by skeptical theis
ts against RNA.
b. Wykstra’s Parent Analogy
Skeptical theists have drawn various analogies in an attempt to highlight the im
plausibility of RNA. The most common analogy, and the one favoured by Wykstra, i
nvolves a comparison between the vision and wisdom of an omniscient being such a
s God and the cognitive capacities of members of the human species. Clearly, the
gap between God’s intellect and ours is immense, and Wykstra (1984: 87-91) compar
es it to the gap between the cognitive abilities of a parent and her one-month-o
ld infant. But if this is the case, then even if there were outweighing goods co
nnected in the requisite way to the instances of suffering appealed to by Rowe,
that we should discern most of these goods is just as likely as that a one-month
-old infant should discern most of her parents’ purposes for those pains they allo
w her to suffer – that is to say, it is not likely at all. Assuming that CORNEA is
correct, Rowe would not then be entitled to claim, for any given instance of ap
parently pointless suffering, that it is indeed pointless. For as the above comp
arison between God’s intellect and the human mind indicates, even if there were ou
tweighing goods served by certain instances of suffering, such goods would be be
yond our ken. What Rowe has failed to see, according to Wykstra, is that “if we th
ink carefully about the sort of being theism proposes for our belief, it is enti
rely expectable – given what we know of our cognitive limits – that the goods by vir
tue of which this Being allows known suffering should very often be beyond our k
en” (1984: 91).
c. Alston’s Analogies
Rowe, like many others, has responded to Wykstra’s Parent Analogy by identifying a
number of relevant disanalogies between a one-month-old infant and our predicam
ent as adult human beings (see Rowe 1996: 275). There are, however, various othe
r analogies that skeptical theists have employed in order to cast doubt on RNA.
Here I will briefly consider a series of analogies that were first formulated by
Alston (1996).
Like Wykstra, Alston (1996: 317) aims to highlight “the absurdity of the claim” that
the fact that we cannot see what justifying reason an omniscient, omnipotent be
ing might have for doing something provides strong support for the supposition t
hat no such reason is available to that being. Alston, however, chooses to steer
clear of the parent-child analogy employed by Wykstra, for he concedes that thi
s contains loopholes that can be exploited in the ways suggested by Rowe.
Alston’s analogies fall into two groups, the first of which attempt to show that t
he insights attainable by finite, fallible human beings are not an adequate indi
cation of what is available by way of reasons to an omniscient, omnipotent being
. Suppose I am a first-year university physics student and I am faced with a the
ory of quantum phenomena, but I struggle to see why the author of the theory dra
ws the conclusions she draws. Does that entitle me to suppose that she has no su
fficient reason for her conclusions? Clearly not, for my inability to discern he
r reasons is only to be expected given my lack of expertise in the subject. Simi
larly, given my lack of training in painting, I fail to see why Picasso arranged
the figures in Guernica as he did. But that does not entitle me to infer that h
e had no sufficient reason for doing so. Again, being a beginner in chess, I fai
l to see any reason why Kasparov made the move he did, but I would be foolish to
conclude that he had no good reason to do so.
Alston applies the foregoing to the noseeum inference from “We cannot see any suff
icient reason for God to permit E1 and E2″ to “God has no sufficient reason to do so
.” In this case, as in the above examples, we are in no position to draw such a co
nclusion for we lack any reason to suppose that we have a sufficient grasp of th
e range of possible reasons open to the other party. Our grasp of the reasons Go
d might have for his actions is thus comparable to the grasp of the neophyte in
the other cases. Indeed, Alston holds that “the extent to which God can envisage r
easons for permitting a given state of affairs exceeds our ability to do so by a
t least as much as Einstein’s ability to discern the reason for a physical theory
exceeds the ability of one ignorant of physics” (1996: 318, emphasis his).
Alston’s second group of analogies seek to show that, in looking for the reasons G
od might have for certain acts or omissions, we are in effect trying to determin
e whether there is a so-and-so in a territory the extent and composition of whic
h is largely unknown to us (or, at least, it is a territory such that we have no
way of knowing the extent to which its constituents are unknown to us). Alston
thus states that Rowe’s noseeum inference
…is like going from “We haven’t found any signs of life elsewhere in the universe” to “The
re isn’t life elsewhere in the universe.” It is like someone who is culturally and g
eographically isolated going from “As far as I have been able to tell, there is no
thing on earth beyond this forest” to “There is nothing on earth beyond this forest.”
Or, to get a bit more sophisticated, it is like someone who reasons “We are unable
 to discern anything beyond the temporal bounds of our universe,” where those boun
ds are the big bang and the final collapse, to “There is nothing beyond the tempor
al bounds of our universe.” (1996: 318)
Just as we lack a map of the relevant “territory” in these cases, we also lack a rel
iable internal map of the diversity of considerations that are available to an o
mniscient being in permitting instances of suffering. But given our ignorance of
the extent, variety, or constitution of the terra incognita, it is surely the b
etter part of wisdom to refrain from drawing any hasty conclusions regarding the
nature of this territory.
Although such analogies may not be open to the same criticisms levelled against
the analogies put forward by Wykstra, they are in the end no more successful tha
n Wykstra’s analogies. Beginning with Alston’s first group of analogies, where a nos
eeum inference is unwarranted due to a lack of expertise, there is typically no
expectation on the part of the neophyte that the reasons held by the other party
(e.g., the physicist’s reasons for drawing conclusion x, Kasparov’s reasons for mak
ing move x in a chess game) would be discernible to her. If you have just begun
to study physics, you would not expect to understand Einstein’s reasons for advanc
ing the special theory of relativity. However, if your five-year-old daughter su
ffered the fate of Sue as depicted in E2, would you not expect a perfectly lovin
g being to reveal his reasons to you for allowing this to happen, or at least to
comfort you by providing you with special assurances that that there is a reaso
n why this terrible evil could not have been prevented? Rowe makes this point qu
ite well:
Being finite beings we can’t expect to know all the goods God would know, any more
than an amateur at chess should expect to know all the reasons for a particular
move that Kasparov makes in a game. But, unlike Kasparov who in a chess match h
as a good reason not to tell us how a particular move fits into his plan to win
the game, God, if he exists, isn’t playing chess with our lives. In fact, since un
derstanding the goods for the sake of which he permits terrible evils to befall
us would itself enable us to better bear our suffering, God has a strong reason
to help us understand those goods and how they require his permission of the ter
rible evils that befall us. (2001b: 157)
There appears, then, to be an obligation on the part of a perfect being to not k
eep his intentions entirely hidden from us. Such an obligation, however, does no
t attach to a gifted chess player or physicist – Kasparov cannot be expected to re
veal his game plan, while a physics professor cannot be expected to make her mat
hematical demonstration in support of quantum theory comprehensible to a high sc
hool physics student.
Similarly with Alston’s second set of analogies, where our inability to map the te
rritory within which to look for x is taken to preclude us from inferring from o
ur inability to find x that there is no x. This may be applicable to cases like
the isolated tribesman’s search for life outside his forest or our search for extr
aterrestrial life, for in such scenarios there is no prior expectation that the
objects of our search are of such a nature that, if they exist, they would make
themselves manifest to us. However, in our search for God’s reasons we are toiling
in a unique territory, one inhabited by a perfectly loving being who, as such,
would be expected to make at least his presence, if not also his reasons for per
mitting evil, (more) transparent to us. This difference in prior expectations un
covers an important disanalogy between the cases Alston considers and cases invo
lving our attempt to discern God’s intentions. Alston’s analogies, therefore, not on
ly fail to advance the case against RNA but also suggest a line of thought in su
pport of RNA. (For further discussion on RNA and divine hiddenness, see Trakakis
(2003); see also Howard-Snyder & Moser (2002).)
4. Building a Theodicy, or Casting Light on the Ways of God
Most critics of Rowe’s evidential argument have thought that the problem with the
argument lies with its factual premise. But what, exactly, is wrong with this pr
emise? According to one popular line of thought, the factual premise can be show
n to be false by identifying goods that we know of that would justify God in per
mitting evil. To do this is to develop a theodicy.
a. What Is a Theodicy?
The primary aim of the project of theodicy may be characterized in John Milton’s c
elebrated words as the attempt to “justify the ways of God to men.” That is to say,
a theodicy aims to vindicate the justice or goodness of God in the face of the e
vil found in the world, and this it attempts to do by offering a reasonable expl
anation as to why God allows evil to abound in his creation.
A theodicy may be thought of as a story told by the theist explaining why God pe
rmits evil. Such a story, however, must be plausible or reasonable in the sense
that it conforms to all of the following:
1. commonsensical views about the world (e.g., that there exist other peopl
e, that there exists a mind-independent world, that much evil exists);
2. widely accepted scientific and historical views (e.g., evolutionary theo
ry), and
3. intuitively plausible moral principles (e.g., generally, punishment shou
ld not be significantly disproportional to the offence committed).
Judged by these criteria, the story of the Fall (understood in a literalist fash
ion) could not be offered as a theodicy. For given the doubtful historicity of A
dam and Eve, and given the problem of harmonizing the Fall with evolutionary the
ory, such an account of the origin of evil cannot reasonably held to be plausibl
e. A similar point could be made about stories that attempt to explain evil as t
he work of Satan and his cohorts.
b. Distinguishing a “Theodicy” from a “Defence”
An important distinction is often made between a defence and a theodicy. A theod
icy is intended to be a plausible or reasonable explanation as to why God permit
s evil. A defence, by contrast, is only intended as a possible explanation as to
why God permits evil. A theodicy, moreover, is offered as a solution to the evi
dential problem of evil, whereas a defence is offered as a solution to the logic
al problem of evil. Here is an example of a defence, which may clarify this dist
inction:
It will be recalled that, according to Mackie, it is logically impossible for th
e following two propositions to be jointly true:
1. God is omnipotent, omniscient, and perfectly good,
2. Evil exists.
Now, consider the following proposition:
1. Every person goes wrong in every possible world.
In other words, every free person created by God would misuse their free will on
at least one occasion, no matter which world (or what circumstances) they were
placed in. This may be highly implausible, or even downright false – but it is, at
least, logically possible. And if (3) is possible, then so is the following pro
position:
1. It was not within God’s power to create a world containing moral good but
no moral evil.
In other words, it is possible that any world created by God that contains some
moral good will also contain some moral evil. Therefore, it is possible for both
(1) and (2) to be jointly true, at least when (2) is said to refer to “moral evil
.” But what about “natural evil”? Well, consider the following proposition:
1. All so-called “natural evil” is brought about by the devious activities of S
atan and his cohorts.
In other words, what we call “natural evil” is actually “moral evil” since it results fr
om the misuse of someone’s free will (in this case, the free will of some evil dem
on). Again, this may be highly implausible, or even downright false – but it is, a
t least, possibly true.
In sum, Mackie was wrong to think that it is logically impossible for both (1) a
nd (2) to be true. For if you conjoin (4) and (5) to (1) and (2), it becomes cle
ar that it is possible that any world created by God would have some evil in it.
(This, of course, is the famous free will defence put forward in Plantinga 1974
: ch.9). Notice that the central claims of this defence – namely, (3), (4), and (5
) – are only held to be possibly true. That’s what makes this a defence. One could n
ot get away with this in a theodicy, for a theodicy must be more than merely pos
sibly true.
c. Sketch of a Theodicy
What kind of theodicy, then, can be developed in response to Rowe’s evidential arg
ument? Are there any goods we know of that would justify God in permitting evils
like E1 and E2? Here I will outline a proposal consisting of three themes that
have figured prominently in the recent literature on the project of theodicy.
(1) Soul-making. Inspired by the thought of the early Church Father, Irenaeus of
Lyon (c.130-c.202 CE), John Hick has put forward in a number of writings, but a
bove all in his 1966 classic Evil and the God of Love, a theodicy that appeals t
o the good of soul-making (see also Hick 1968, 1977, 1981, 1990). According to H
ick, the divine intention in relation to humankind is to bring forth perfect fin
ite personal beings by means of a “vale of soul-making” in which humans may transcen
d their natural self-centredness by freely developing the most desirable qualiti
es of moral character and entering into a personal relationship with their Maker
. Any world, however, that makes possible such personal growth cannot be a hedon
istic paradise whose inhabitants experience a maximum of pleasure and a minimum
of pain. Rather, an environment that is able to produce the finest characteristi
cs of human personality – particularly the capacity to love – must be one in which “th
ere are obstacles to be overcome, tasks to be performed, goals to be achieved, s
etbacks to be endured, problems to be solved, dangers to be met” (Hick 1966: 362).
A soul-making environment must, in other words, share a good deal in common wit
h our world, for only a world containing great dangers and risks, as well as the
genuine possibility of failure and tragedy, can provide opportunities for the d
evelopment of virtue and character. A necessary condition, however, for this dev
elopmental process to take place is that humanity be situated at an “epistemic dis
tance” from God. On Hick’s view, in other words, if we were initially created in the
direct presence of God we could not freely come to love and worship God. So as
to preserve our freedom in relation to God, the world must be created religiousl
y ambiguous or must appear, to some extent at least, as if there were no God. An
d evil, of course, plays an important role in creating the desired epistemic dis
tance.
(2) Free will. The appeal to human freedom, in one guise or another, constitutes
an enduring theme in the history of theodicy. Typically, the kind of freedom th
at is invoked by the theodicist is the libertarian sort, according to which I am
free with respect to a particular action at time t only if the action is not de
termined by all that happened or obtained before t and all the causal laws there
are in such a way that the conjunction of the two (the past and the laws) logic
ally entails that I perform the action in question. My mowing the lawn, for inst
ance, constitutes a voluntary action only if, the state of the universe (includi
ng my beliefs and desires) and laws of nature being just as they were immediatel
y preceding my decision to mow the lawn, I could have chosen or acted otherwise
than I in fact did. In this sense, the acts I perform freely are genuinely “up to
me” – they are not determined by anything external to my will, whether these be caus
al laws or even God. And so it is not open to God to cause or determine just wha
t actions I will perform, for if he does so those actions could not be free. Fre
edom and determinism are incompatible.
The theodicist, however, is not so much interested in libertarian freedom as in
libertarian freedom of the morally relevant kind, where this consists of the fre
edom to choose between good and evil courses of action. The theodicist’s freedom,
moreover, is intended to be morally significant, not only providing one with the
capacity to bring about good and evil, but also making possible a range of acti
ons that vary enormously in moral worth, from great and noble deeds to horrific
evils.
Armed therefore with such a conception of freedom, the free will theodicist proc
eeds to explain the existence of moral evil as a consequence of the misuse of ou
r freedom. This, however, means that responsibility for the existence of moral e
vil lies with us, not with God. Of course, God is responsible for creating the c
onditions under which moral evil could come into existence. But it was not inevi
table that human beings, if placed in those conditions, would go wrong. It was n
ot necessary, in other words, that humans would misuse their free will, although
 this always was a possibility and hence a risk inherent in God’s creation of free
creatures. The free will theodicist adds, however, that the value of free will
(and the goods it makes possible) is so great as to outweigh the risk that it ma
y be misused in various ways.
(3) Heavenly bliss. Theodicists sometimes draw on the notion of a heavenly after
life to show that evil, particularly horrendous evil, only finds its ultimate ju
stification or redemption in the life to come. Accounts of heaven, even within t
he Christian tradition, vary widely. But one common feature in these accounts th
at is relevant to the theodicist’s task is the experience of complete felicity for
eternity brought about by intimate and loving communion with God. This good, as
we saw, plays an important role in Hick’s theodicy, and it also finds a central p
lace in Marilyn Adams’ account of horrendous evil.
Adams (1986: 262-63, 1999: 162-63) notes that, on the Christian world-view, the
direct experience of “face-to-face” intimacy with God is not only the highest good w
e can aspire to enjoy, but is also an incommensurable good – more precisely, it is
incommensurable with respect to any merely temporal evils or goods. As the apos
tle Paul put it, “our present sufferings are not worth comparing with the glory th
at will be revealed in us” (Rom 8:18, NIV; compare 2 Cor 4:17). This glorification
to be experienced in heaven, according to Adams, vindicates God’s justice and lov
e toward his creatures. For the experience of the beatific vision outweighs any
evil, even evil of the horrendous variety, that someone may suffer, thus ensurin
g a balance of good over evil in the sufferer’s life that is overwhelmingly favour
able. But as Adams points out, “strictly speaking, there will be no balance to be
struck” (1986: 263, emphasis hers), since the good of the vision of God is incomme
nsurable with respect to one’s participation in any temporal or created evils. And
so an everlasting, post-mortem beatific vision of God would provide anyone who
experienced it with good reason for considering their life – in spite of any horro
rs it may have contained – as a great good, thus removing any grounds of complaint
against God.
Bringing these three themes together, a theodicy can be developed with the aim o
f explaining and justifying God’s permission of evil, even evil of the horrendous
variety. To illustrate how this may be done, I will concentrate on Rowe’s E2 and t
he Holocaust, two clear instances of horrendous moral evil.
Notice that these two evils clearly involve a serious misuse of free will on beh
alf of the perpetrators. We could, therefore, begin by postulating God’s endowment
of humans with morally significant free will as the first good that is served b
y these evils. That is to say, God could not prevent the terrible suffering and
death endured by Sue and the millions of Holocaust victims while at the same tim
e creating us without morally significant freedom – the freedom to do both great e
vil and great good. In addition, these evils may provide an opportunity for soul
-making – in many cases, however, the potential for soul-making would not extend t
o the victim but only to those who cause or witness the suffering. The phenomeno
n of “jailhouse conversions,” for example, testifies to the fact that even horrendou
s evil may occasion the moral transformation of the perpetrator. Finally, to ade
quately compensate the victims of these evils we may introduce the doctrine of h
eaven. Postmortem, the victims are ushered into a relation of beatific intimacy
with God, an incommensurable good that “redeems” their past participation in horrors
. For the beatific vision in the afterlife not only restores value and meaning t
o the victim’s life, but also provides them with the opportunity to endorse their
life (taken as a whole) as worthwhile.
Does this theodicy succeed in exonerating God? Various objections could, of cour
se, be raised against such a theodicy. One could, for example, question the inte
lligibility or empirical adequacy of the underlying libertarian notion of free w
ill (see, e.g., Pereboom 2001: 38-88). Or one might follow Tooley (1980:373-75)
and Rowe (1996: 279-81, 2001a: 135-36) in thinking that, just as we have a duty
to curtail another person’s exercise of free will when we know that they will use
their free will to inflict considerable suffering on an innocent (or undeserving
) person, so too does God have a duty of this sort. On this view, a perfectly go
od God would have intervened to prevent us from misusing our freedom to the exte
nt that moral evil, particularly moral evil of the horrific kind, would either n
ot occur at all or occur on a much more infrequent basis. Finally, how can the a
bove theodicy be extended to account for natural evil? Various proposals have be
en offered here, the most prominent of which are: Hick’s view that natural evil pl
ays an essential part in the “soul-making” process; Swinburne’s “free will theodicy for
natural evil” – the idea, roughly put, is that free will cannot be had without the k
nowledge of how to bring about evil (or prevent its occurrence), and since this
knowledge of how to cause evil can only be had through prior experience with nat
ural evil, it follows that the existence of natural evil is a logically necessar
y condition for the exercise of free will (see Swinburne 1978, 1987: 149-67, 199
1: 202-214, 1998: 176-92); and “natural law theodicies,” such as that developed by R
eichenbach (1976, 1982: 101-118), according to which the natural evils that befa
ll humans and animals are the unavoidable by-products of the outworking of the n
atural laws governing God’s creation.
5. Further Responses to the Evidential Problem of Evil
Let’s suppose that Rowe’s evidential argument from evil succeeds in providing strong
evidence in support of the claim that there does not exist an omnipotent, omnis
cient, wholly good being. What follows from this? In particular, would a theist
who finds its impossible to fault Rowe’s argument be obliged to give up her theism
? Not necessarily, for at least two further options would be available to such a
theist.
Firstly, the theist may agree that Rowe’s argument provides some evidence against
theism, but she may go on to argue that there is independent evidence in support
of theism which outweighs the evidence against theism. In fact, if the theist t
hinks that the evidence in support of theism is quite strong, she may employ wha
t Rowe (1979: 339) calls “the G.E. Moore shift” (compare Moore 1953: ch.6). This inv
olves turning the opponent’s argument on its head, so that one begins by denying t
he very conclusion of the opponent’s argument. The theist’s counter-argument would t
hen proceed as follows:
(not-3) There exists an omnipotent, omniscient, wholly good being.
(2) An omniscient, wholly good being would prevent the occurrence of any int
ense suffering it could, unless it could not do so without thereby losing some g
reater good or permitting some evil equally bad or worse.
(not-1) (Therefore) It is not the case that there exist instances of horrendous
evil which an omnipotent, omniscient being could have prevented without thereby
losing some greater good or permitting some evil equally bad or worse.
Although this strategy has been welcomed by many theists as an appropriate way o
f responding to evidential arguments from evil (e.g., Mavrodes 1970: 95-97, Evan
s 1982: 138-39, Davis 1987: 86-87, Basinger 1996: 100-103) – indeed, it is conside
red by Rowe to be “the theist’s best response” (1979: 339) – it is deeply problematic in
a way that is often overlooked. The G.E. Moore shift, when employed by the thei
st, will be effective only if the grounds for accepting not-(3) [the existence o
f the theistic God] are more compelling than the grounds for accepting not-(1) [
the existence of gratuitous evil]. The problem here is that the kind of evidence
that is typically invoked by theists in order to substantiate the existence of
God – e.g., the cosmological and design arguments, appeals to religious experience
– does not even aim to establish the existence of a perfectly good being, or else
, if it does have such an aim, it faces formidable difficulties in fulfilling it
. But if this is so, then the theist may well be unable to offer any evidence at
all in support of not-(3), or at least any evidence of a sufficiently strong or
cogent nature in support of not-(3). The G.E. Moore shift, therefore, is not as
straightforward a strategy as it initially seems.
Secondly, the theist who accepts Rowe’s argument may claim that Rowe has only show
n that one particular version of theism – rather than every version of theism – need
s to be rejected. A process theist, for example, may agree with Rowe that there
is no omnipotent being, but would add that God, properly understood, is not omni
potent, or that God’s power is not as unlimited as is usually thought (see, e.g.,
Griffin 1976, 1991). An even more radical approach would be to posit a “dark side” i
n God and thus deny that God is perfectly good. Theists who adopt this approach
(e.g., Blumenthal 1993, Roth 2001) would also have no qualms with the conclusion
of Rowe’s argument.
There are at least two problems with this second strategy. Firstly, Rowe’s argumen
t is only concerned with the God of orthodox theism as described in Section I.1
above, not the God of some other version of theism. And so objections drawn from
non-orthodox forms of theism fail to engage with Rowe’s argument (although such o
bjections may be useful in getting us to reconsider the traditional understandin
g of God). A second problem concerns the worship-worthiness of the sort of deity
being proposed. For example, would someone who is not wholly good and capable o
f evil be fit to be the object of our worship, total devotion and unconditional
commitment? Similarly, why place complete trust in a God who is not all-powerful
and hence not in full control of the world? (To be sure, even orthodox theists
will place limits on God’s power, and such limits on divine power may go some way
towards explaining the presence of evil in the world. But if God’s power, or lack
thereof, is offered as the solution to the problem of evil – so that the reason wh
y God allows evil is because he doesn’t have the power to prevent it from coming i
nto being – then we are faced with a highly impotent God who, insofar as he is awa
re of the limitations in his power, may be considered reckless for proceeding wi
th creation.)
6. Conclusion
Evidential arguments from evil, such as those developed by William Rowe, purport
to show that, grounds for belief in God aside, the existence of evil renders at
heism more reasonable than theism. What verdict, then, can be reached regarding
such arguments? A brief answer to this question may be provided by way of an ove
rview of the foregoing investigation.
Firstly, as was argued in Section II, the “open theist” response to Rowe’s theological
premise either runs the risk of diminishing confidence in God or else is entire
ly compatible with the theological premise. Secondly, the “sceptical theist” objecti
on to Rowe’s inference from inscrutable evil to pointless evil was examined in Sec
tion III and was found to be inadequately supported. Thirdly, various theodical
options were canvassed in Section IV as a possible way of refuting Rowe’s factual
premise, and it was found that a theodicy that appeals to the goods of free will
, soul-making, and a heavenly afterlife may go some way in accounting for the ex
istence of moral evil. Such a theodicy, however, raises many further questions r
elating to the existence of natural evil and the existence of so much horrendous
moral evil. And finally, as argued in Section V, the strategy of resorting to t
he “G.E. Moore shift” faces the daunting task of furnishing evidence in support of t
he existence of a perfect being; while resorting to a non-orthodox conception of
God dissolves the problem of evil at the cost of corroding religiously signific
ant attitudes and practices such as the love and worship of God.
On the basis of these results it can be seen that Rowe’s argument has a strongly r
esilient character, successfully withstanding many of the objections raised agai
nst it. Much more, of course, can be said both in support of and against Rowe’s ca
se for atheism. Although it might therefore be premature to declare any one side
to the debate victorious, it can be concluded that, at the very least, Rowe’s evi
dential argument is not as easy to refute as is often presumed.

Logical Problem of Evil

The existence of evil and suffering in our world seems to pose a serious challen
ge to belief in the existence of a perfect God. If God were all-knowing, it seem
s that God would know about all of the horrible things that happen in our world.
If God were all-powerful, God would be able to do something about all of the ev
il and suffering. Furthermore, if God were morally perfect, then surely God woul
d want to do something about it. And yet we find that our world is filled with c
ountless instances of evil and suffering. These facts about evil and suffering
seem to conflict with the orthodox theist claim that there exists a perfectly go
od God. The challenged posed by this apparent conflict has come to be known as t
he problem of evil.
This article addresses one form of that problem that is prominent in recent phil
osophical discussions–that the conflict that exists between the claims of orthodox
 theism and the facts about evil and suffering in our world is a logical one. Th
is is the “logical problem of evil.”
The article clarifies the nature of the logical problem of evil and considers va
rious theistic responses to the problem. Special attention is given to the free
will defense, which has been the most widely discussed theistic response to the
logical problem of evil.
Table of Contents
1. Introducing the Problem
2. Logical Consistency
3. Logical Consistency and the Logical Problem of Evil
4. Plantinga’s Free Will Defense
5. Divine Omnipotence and the Free Will Defense
6. An Objection: Free Will and Natural Evil
7. Evaluating the Free Will Defense
8. Was Plantinga’s Victory Too Easy?
9. Other Responses to the Logical Problem of Evil
10. Problems with the Free Will Defense
11. References and Further Reading
1. References
2. Further Reading
1. Introducing the Problem
Journalist and best-selling author Lee Strobel commissioned George Barna, the pu
blic-opinion pollster, to conduct a nationwide survey. The survey included the q
uestion “If you could ask God only one question and you knew he would give you an
answer, what would you ask?” The most common response, offered by 17% of those who
could think of a question was “Why is there pain and suffering in the world?” (Stro
bel 2000, p. 29). If God is all-powerful, all-knowing and perfectly good, why do
es he let so many bad things happen? This question raises what philosophers call
“the problem of evil.”
It would be one thing if the only people who suffered debilitating diseases or t
ragic losses were the likes of Adolf Hitler, Joseph Stalin or Osama Bin Laden. A
s it is, however, thousands of good-hearted, innocent people experience the rava
ges of violent crime, terminal disease, and other evils. Michael Peterson (1998,
p. 1) writes,
Something is dreadfully wrong with our world. An earthquake kills hundreds in Pe
ru. A pancreatic cancer patient suffers prolonged, excruciating pain and dies. A
pit bull attacks a two-year-old child, angrily ripping his flesh and killing hi
m. Countless multitudes suffer the ravages of war in Somalia. A crazed cult lead
er pushes eighty-five people to their deaths in Waco, Texas. Millions starve and
die in North Korea as famine ravages the land. Horrible things of all kinds hap
pen in our world—and that has been the story since the dawn of civilization.
Peterson (1998, p. 9) claims that the problem of evil is a kind of “moral protest.”
In asking “How could God let this happen?” people are often claiming “It’s not fair that
God has let this happen.” Many atheists try to turn the existence of evil and suf
fering into an argument against the existence of God. They claim that, since the
re is something morally problematic about a morally perfect God allowing all of
the evil and suffering we see, there must not be a morally perfect God after all
. The popularity of this kind of argument has led Hans Küng (1976, p. 432) to call
the problem of evil “the rock of atheism.” This essay examines one form the argumen
t from evil has taken, which is known as “the logical problem of evil.”
In the second half of the twentieth century, atheologians (that is, persons who
try to prove the non-existence of God) commonly claimed that the problem of evil
was a problem of logical inconsistency. J. L. Mackie (1955, p. 200), for exampl
e, claimed,
Here it can be shown, not that religious beliefs lack rational support, but that
they are positively irrational, that several parts of the essential theological
doctrine are inconsistent with one another.
H. J. McCloskey (1960, p. 97) wrote,
Evil is a problem, for the theist, in that a contradiction is involved in the fa
ct of evil on the one hand and belief in the omnipotence and omniscience of God
on the other.
Mackie and McCloskey can be understood as claiming that it is impossible for all
of the following statements to be true at the same time:
(1) God is omnipotent (that is, all-powerful).
(2) God is omniscient (that is, all-knowing).
(3) God is perfectly good.
(4) Evil exists.
Any two or three of them might be true at the same time; but there is no way tha
t all of them could be true. In other words, (1) through (4) form a logically in
consistent set. What does it mean to say that something is logically inconsisten
t?
(5) A set of statements is logically inconsistent if and only if: (a) that set i
ncludes a direct contradiction of the form “p & not-p”; or (b) a direct contradictio
n can be deduced from that set.
None of the statements in (1) through (4) directly contradicts any other, so if
the set is logically inconsistent, it must be because we can deduce a contradict
ion from it. This is precisely what atheologians claim to be able to do.
Atheologians claim that a contradiction can easily be deduced from (1) through (
4) once we think through the implications of the divine attributes cited in (1)
through (3). They reason as follows:
(6) If God is omnipotent, he would be able to prevent all of the evil and suffer
ing in the world.
(7) If God is omniscient, he would know about all of the evil and suffering in t
he world and would know how to eliminate or prevent it.
(8) If God is perfectly good, he would want to prevent all of the evil and suffe
ring in the world.
Statements (6) through (8) jointly imply that if the perfect God of theism reall
y existed, there would not be any evil or suffering. However, as we all know, ou
r world is filled with a staggering amount of evil and suffering. Atheologians c
laim that, if we reflect upon (6) through (8) in light of the fact of evil and s
uffering in our world, we should be led to the following conclusions:
(9) If God knows about all of the evil and suffering in the world, knows how to
eliminate or prevent it, is powerful enough to prevent it, and yet does not prev
ent it, he must not be perfectly good.
(10) If God knows about all of the evil and suffering, knows how to eliminate or
prevent it, wants to prevent it, and yet does not do so, he must not be all- po
werful.
(11) If God is powerful enough to prevent all of the evil and suffering, wants t
o do so, and yet does not, he must not know about all of the suffering or know h
ow to eliminate or prevent it—that is, he must not be all-knowing.
From (9) through (11) we can infer:
(12) If evil and suffering exist, then God is either not omnipotent, not omnisci
ent, or not perfectly good.
Since evil and suffering obviously do exist, we get:
(13) God is either not omnipotent, not omniscient, or not perfectly good.
Putting the point more bluntly, this line of argument suggests that—in light of th
e evil and suffering we find in our world—if God exists, he is either impotent, ig
norant or wicked. It should be obvious that (13) conflicts with (1) through (3)
above. To make the conflict more clear, we can combine (1), (2) and (3) into the
following single statement.
(14) God is omnipotent, omniscient and perfectly good.
There is no way that (13) and (14) could both be true at the same time. These st
atements are logically inconsistent or contradictory.
Statement (14) is simply the conjunction of (1) through (3) and expresses the ce
ntral belief of classical theism. However, atheologians claim that statement ( 1
3) can also be derived from (1) through (3). [Statements (6) through (12) purpor
t to show how this is done.] (13) and (14), however, are logically contradictory
. Because a contradiction can be deduced from statements (1) through (4) and bec
ause all theists believe (1) through (4), atheologians claim that theists have l
ogically inconsistent beliefs. They note that philosophers have always believed
it is never rational to believe something contradictory. So, the existence of ev
il and suffering makes theists’ belief in the existence of a perfect God irrationa
l.
Can the believer in God escape from this dilemma? In his best-selling book When
Bad Things Happen to Good People, Rabbi Harold Kushner (1981) offers the followi
ng escape route for the theist: deny the truth of (1). According to this proposa
l, God is not ignoring your suffering when he doesn’t act to prevent it because—as a
n all-knowing God—he knows about all of your suffering. As a perfectly good God, h
e also feels your pain. The problem is that he can’t do anything about it because
he’s not omnipotent. According to Kushner’s portrayal, God is something of a kind-he
arted wimp. He’d like to help, but he doesn’t have the power to do anything about ev
il and suffering. Denying the truth of either (1), (2), (3) or ( 4) is certainly
one way for the theist to escape from the logical problem of evil, but it would
not be a very palatable option to many theists. In the remainder of this essay,
we will examine some theistic responses to the logical problem of evil that do
not require the abandonment of any central tenet of theism.
2. Logical Consistency
Theists who want to rebut the logical problem of evil need to find a way to show
that (1) through (4)—perhaps despite initial appearances—are consistent after all.
We said above that a set of statements is logically inconsistent if and only if
that set includes a direct contradiction or a direct contradiction can be deduce
d from that set. That means that a set of statements is logically consistent if
and only if that set does not include a direct contradiction and a direct contra
diction cannot be deduced from that set. In other words,
(15) A set of statements is logically consistent if and only if it is possible f
or all of them to be true at the same time.
Notice that (15) does not say that consistent statements must actually be true a
t the same time. They may all be false or some may be true and others false. Con
sistency only requires that it be possible for all of the statements to be true
(even if that possibility is never actualized). (15) also doesn’t say anything abo
ut plausibility. It does not require the joint of a consistent set of statements
to be plausible. It may be exceedingly unlikely or improbable that a certain se
t of statements should all be true at the same time. But improbability is not th
e same thing as impossibility. As long as there is nothing contradictory about t
heir conjunction, it will be possible (even if unlikely) for them all to be true
at the same time.
This brief discussion allows us to see that the atheological claim that statemen
ts (1) through (4) are logically inconsistent is a rather strong one. The atheol
ogian is maintaining that statements (1) through (4) couldn’t possibly all be true
at the same time. In other words,
(16) It is not possible for God and evil to co-exist.
The logical problem of evil claims that God’s omnipotence, omniscience and supreme
goodness would completely rule out the possibility of evil and that the existen
ce of evil would do the same for the existence of a supreme being.
3. Logical Consistency and the Logical Problem of Evil
How might a theist go about demonstrating that (16) is false? Some theists sugge
st that perhaps God has a good reason for allowing the evil and suffering that h
e does. Not just any old reason can justify God’s allowing all of the evil and suf
fering we see. Mass murderers and serial killers typically have reasons for why
they commit horrible crimes, but they do not have good reasons. It’s only when peo
ple have morally good reasons that we excuse or condone their behavior. Philosop
hers of religion have called the kind of reason that could morally justify God’s a
llowing evil and suffering a “morally sufficient reason.”
Consider the following statement.
(17) It is possible that God has a morally sufficient reason for allowing evil.
If God were to have a morally sufficient reason for allowing evil, would it be p
ossible for God to be omnipotent, omniscient, perfectly good, and yet for there
to be evil and suffering? Many theists answer “Yes.” If (17) were true, (9) through
(12) would have to be modified to read:
(9′) If God knows about all of the evil and suffering in the world, knows how to e
liminate or prevent it, is powerful enough to prevent it, and yet does not preve
nt it, he must not be perfectly good—unless he has a morally sufficient reason for
allowing evil.
(10′) If God knows about all of the evil and suffering, knows how to eliminate or
prevent it, wants to prevent it, and yet does not do so, he must not be all-powe
rful—unless he has a morally sufficient reason for allowing evil.
(11′) If God is powerful enough to prevent all of the evil and suffering, wants to
do so, and yet does not, he must not know about all of the suffering or know ho
w to eliminate or prevent it (that is, he must not be all-knowing)—unless he has a
morally sufficient reason for allowing evil.
(12′) If evil and suffering exist, then either: a) God is not omnipotent, not omni
scient, or not perfectly good; or b) God has a morally sufficient reason for all
owing evil.
From (9′) through (12′), it is not possible to conclude that God does not exist. The
most that can be concluded is that either God does not exist or God has a moral
ly sufficient reason for allowing evil. So, some theists suggest that the real q
uestion behind the logical problem of evil is whether (17) is true.
If it is possible that God has a morally sufficient reason for allowing evil and
suffering to occur, then the logical problem of evil fails to prove the non-exi
stence of God. If, however, it is not possible that God has a morally sufficient
reason for allowing evil, then it seems that (13) would be true: God is either
not omnipotent, not omniscient, or not perfectly good.
An implicit assumption behind this part of the debate over the logical problem o
f evil is the following:
(18) It is not morally permissible for God to allow evil and suffering to occur
unless he has a morally sufficient reason for doing so.
Is (18) correct? Many philosophers think so. It is difficult to see how a God wh
o allowed bad things to happen just for the heck of it could be worthy of revere
nce, faith and worship. If God had no morally sufficient reason for allowing evi
l, then if we made it to the pearly gates some day and asked God why he allowed
so many bad things to happen, he would simply have to shrug his shoulders and sa
y “There was no reason or point to all of that suffering you endured. I just felt
like letting it happen.” This callous image of God is difficult to reconcile with
orthodox theism’s portrayal of God as a loving Father who cares deeply about his c
reation. (18), combined with the assumption that God does not have a morally suf
ficient reason for allowing evil, yields
(19) God is doing something morally inappropriate or blameworthy in allowing evi
l to occur,
and
(20) If God is doing something morally inappropriate or blameworthy, then God is
not perfectly good.
If (19) and (20) are true, then the God of orthodox theism does not exist.
What would it look like for God to have a morally sufficient reason for allowing
evil? Let’s first consider a down-to-earth example of a morally sufficient reason
a human being might have before moving on to the case of God. Suppose a gossipy
neighbor were to tell you that Mrs. Jones just allowed someone to inflict unwan
ted pain upon her child. Your first reaction to this news might be one of horror
. But once you find out that the pain was caused by a shot that immunized Mrs. J
ones’ infant daughter against polio, you would no longer view Mrs. Jones as a dang
er to society. Generally, we believe the following moral principle to be true.
(21) Parents should not inflict unwanted pain upon their children.
In the immunization case, Mrs. Jones has a morally sufficient reason for overrid
ing or suspending this principle. A higher moral duty—namely, the duty of protecti
ng the long-term health of her child—trumps the lesser duty expressed by (21). If
God has a morally sufficient reason for allowing evil and suffering, theists cla
im, it will probably look something like Mrs. Jones’.
4. Plantinga’s Free Will Defense
What might God’s reason be for allowing evil and suffering to occur? Alvin Plantin
ga (1974, 1977) has offered the most famous contemporary philosophical response
to this question. He suggests the following as a possible morally sufficient rea
son:
(MSR1) God’s creation of persons with morally significant free will is something o
f tremendous value. God could not eliminate much of the evil and suffering in th
is world without thereby eliminating the greater good of having created persons
with free will with whom he could have relationships and who are able to love on
e another and do good deeds.
(MSR1) claims that God allows some evils to occur that are smaller in value than
a greater good to which they are intimately connected. If God eliminated the ev
il, he would have to eliminate the greater good as well. God is pictured as bein
g in a situation much like that of Mrs. Jones: she allowed a small evil (the pai
n of a needle) to be inflicted upon her child because that pain was necessary fo
r bringing about a greater good (immunization against polio). Before we try to d
ecide whether (MSR1) can justify God in allowing evil and suffering to occur, so
me of its key terms need to be explained.
To begin with, (MSR1) presupposes the view of free will known as “libertarianism”:
(22) Libertarianism=df the view that a person is free with respect to a given ac
tion if and only if that person is both free to perform that action and free to
refrain from performing that action; in other words, that person is not determin
ed to perform or refrain from that action by any prior causal forces.
Although the term “libertarianism” isn’t exactly a household name, the view it express
es is commonly taken to be the average person’s view of free will. It is the view
that causal determinism is false, that—unlike robots or other machines—we can make c
hoices that are genuinely free.
According to Plantinga, libertarian free will is a morally significant kind of f
ree will. An action is morally significant just when it is appropriate to evalua
te that action from a moral perspective (for example, by ascribing moral praise
or blame). Persons have morally significant free will if they are able to perfor
m actions that are morally significant. Imagine a possible world where God creat
es creatures with a very limited kind of freedom. Suppose that the persons in th
is world can only choose good options and are incapable of choosing bad options.
So, if one of them were faced with three possible courses of action—two of which
were morally good and one of which was morally bad—this person would not be free w
ith respect to the morally bad option. That is, that person would not be able to
choose any bad option even if they wanted to. Our hypothetical person does, how
ever, have complete freedom to decide which of the two good courses of action to
take. Plantinga would deny that any such person has morally significant free wi
ll. People in this world always perform morally good actions, but they deserve n
o credit for doing so. It is impossible for them to do wrong. So, when they do p
erform right actions, they should not be praised. It would be ridiculous to give
moral praise to a robot for putting your soda can in the recycle bin rather tha
n the trash can, if that is what it was programmed to do. Given the program runn
ing inside the robot and its exposure to an empty soda can, it’s going to take the
can to the recycle bin. It has no choice about the matter. Similarly, the peopl
e in the possible world under consideration have no choice about being good. Sin
ce they are pre-programmed to be good, they deserve no praise for it.
According to Plantinga, people in the actual world are free in the most robust s
ense of that term. They are fully free and responsible for their actions and dec
isions. Because of this, when they do what is right, they can properly be praise
d. Moreover, when they do wrong, they can be rightly blamed or punished for thei
r actions.
It is important to note that (MSR1) directly conflicts with a common assumption
about what kind of world God could have created. Many atheologians believe that
God could have created a world that was populated with free creatures and yet di
d not contain any evil or suffering. Since this is something that God could have
done and since a world with free creatures and no evil is better than a world w
ith free creatures and evil, this is something God should have done. Since he di
d not do so, God did something blameworthy by not preventing or eliminating evil
and suffering (if indeed God exists at all). In response to this charge, Planti
nga maintains that there are some worlds God cannot create. In particular, he ca
nnot do the logically impossible. (MSR1) claims that God cannot get rid of much
of the evil and suffering in the world without also getting rid of morally signi
ficant free will. (The question of whether God’s omnipotence is compatible with th
e claim that God cannot do the logically impossible will be addressed below.)
Consider the following descriptions of various worlds. We need to determine whic
h ones describe worlds that are logically possible and which ones describe impos
sible worlds. The worlds described will be possible if the descriptions of those
worlds are logically consistent. If the descriptions of those worlds are incons
istent or contradictory, the worlds in question will be impossible.
W1: (a) God creates persons with morally significant free will;
(b) God does not causally determine people in every situation to choose
what is right and to avoid what is wrong; and
(c) There is evil and suffering in W1.
W2: (a) God does not create persons with morally significant free will;
(b) God causally determines people in every situation to choose what is
right and to avoid what is wrong; and
(c) There is no evil or suffering in W2.
W3: (a) God creates persons with morally significant free will;
(b) God causally determines people in every situation to choose what is
right and to avoid what is wrong; and
(c) There is no evil or suffering in W3.
W4: (a) God creates persons with morally significant free will;
(b) God does not causally determine people in every situation to choose
what is right and to avoid what is wrong; and
(c) There is no evil or suffering in W4.
Let’s figure out which of these worlds are possible. Is W1 possible? Yes. In fact,
on the assumption that God exists, it seems to describe the actual world. Peopl
e have free will in this world and there is evil and suffering. God has obviousl
y not causally determined people in every situation to choose what is right and
to avoid what is wrong because there would be no evil or suffering if he had. So
, W1 is clearly possible.
What about W2? Granting Plantinga’s assumption that human beings are genuinely fre
e creatures, the first thing to notice about W2 is that you and I would not exis
t in such a world. We are creatures with morally significant free will. If you t
ook away our free will, we would no longer be the kinds of creatures we are. We
would not be human in that world. Returning to the main issue, there does not se
em to be anything impossible about God causally determining people in every situ
ation to choose what is right and to avoid what is wrong. It seems clearly possi
ble that whatever creatures God were to make in such a world would not have mora
lly significant free will and that there would be no evil or suffering. W2, then
, is also possible.
Now let’s consider the philosophically more important world W3. Is W3 possible? Pl
antinga says, “No.” Parts (a) and (b) of the description of W3 are, he claims, logic
ally inconsistent. In W3 God causally determines people in every situation to ch
oose what is right and to avoid what is wrong. People in this world couldn’t do mo
rally bad things if they wanted to. And yet part of what it means for creatures
to have morally significant free will is that they can do morally bad things whe
never they want to. Think about what it would be like to live in W3. If you want
ed to tell a lie, you would not be able to do so. Causal forces beyond your cont
rol would make you tell the truth on every occasion. You would also be physicall
y incapable of stealing your neighbor’s belongings. In fact, since W3 is a world w
ithout evil of any kind and since merely wanting to lie or steal is itself a bad
thing, the people in W3 would not even be able to have morally bad thoughts or
desires. If God is going to causally determine people in every situation to choo
se what is right and to avoid what is wrong in W3, there is no way that he could
allow them to be free in a morally significant sense. Peterson (1998, p. 39) wr
ites,
if a person is free with respect to an action A, then God does not bring it abou
t or cause it to be the case that she does A or refrains from doing A. For if Go
d brings it about or causes it to be the case in any manner whatsoever that the
person either does A or does not do A, then that person is not really free.
God can’t have it both ways. He can create a world with free creatures or he can c
ausally determine creatures to choose what is right and to avoid what is wrong e
very time; but he can’t do both. God can forcibly eliminate evil and suffering (as
in W2) only at the cost of getting rid of free will.
The fact that W3 is impossible is centrally important to Plantinga’s Free Will Def
ense. Atheologians, as we saw above, claim that God is doing something morally b
lameworthy by allowing evil and suffering to exist in our world. They charge tha
t a good God would and should eliminate all evil and suffering. The assumption b
ehind this charge is that, in so doing, God could leave human free will untouche
d. Plantinga claims that when we think through what robust free will really amou
nts to, we can see that atheologians are (unbeknownst to themselves) asking God
to do the logically impossible. Being upset that God has not done something that
is logically impossible is, according to Plantinga, misguided. He might say, “Of
course he hasn’t done that. It’s logically impossible!” As we will see in section V be
low, Plantinga maintains that divine omnipotence involves an ability to do anyth
ing that is logically possible, but it does not include the ability to do the lo
gically impossible.
Consider W4. Is it possible? Yes! Most people are tempted to answer “No” when first
exposed to this description, but think carefully about it. Although there is no
evil and suffering in this world, it is not because God causally determines peop
le in every situation to choose what is right and to avoid what is wrong. In thi
s world God has given creatures morally significant free will without any string
s attached. If there is nothing bad in this world, it can only be because the fr
ee creatures that inhabit this world have—by their own free will—always chosen to do
the right thing. Is this kind of situation really possible? Yes. Something is l
ogically possible just when it can be conceived without contradiction. There is
nothing contradictory about supposing that there is a possible world where free
creatures always make the right choices and never go wrong. Of course, it’s highly
improbable, given what we know about human nature. But improbability and imposs
ibility, as we said above, are two different things. In fact, according to the J
udeo-Christian story of Adam and Eve, it was God’s will that significantly free hu
man beings would live in the Garden of Eden and always obey God’s commands. If Ada
m and Eve had followed God’s plan, then W4 would have been the actual world.
It is important to note certain similarities between W1 and W4. Both worlds are
populated by creatures with free will and in neither world does God causally det
ermine people to always choose what is right and to avoid what is wrong. The onl
y difference is that, in W1, the free creatures choose to do wrong at least some
of the time, and in W4, the free creatures always make morally good decisions.
In other words, whether there is immorality in either one of these worlds depend
s upon the persons living in these worlds—not upon God. According to Plantinga’s Fre
e Will Defense, there is evil and suffering in this world because people do immo
ral things. People deserve the blame for the bad things that happen—not God. Plant
inga (1974, p. 190) writes,
The essential point of the Free Will Defense is that the creation of a world con
taining moral good is a cooperative venture; it requires the uncoerced concurren
ce of significantly free creatures. But then the actualization of a world W cont
aining moral good is not up to God alone; it also depends upon what the signific
antly free creatures of W would do.
Atheist philosophers such as Anthony Flew and J. L. Mackie have argued that an o
mnipotent God should be able to create a world containing moral good but no mora
l evil. As Flew (1955, p. 149) put it, “If there is no contradiction here then Omn
ipotence might have made a world inhabited by perfectly virtuous people.” Mackie (
1955, p. 209) writes,
If God has made men such that in their free choices they sometimes prefer what i
s good and sometimes what is evil, why could he not have made men such that they
always freely choose the good? If there is no logical impossibility in a man’s ch
oosing the good on one, or on several occasions, there cannot be a logical impos
sibility in his freely choosing the good on every occasion. God was not, then, f
aced with a choice between making innocent automata and making beings who, in ac
ting freely, would sometimes go wrong: there was open to him the obviously bette
r possibility of making beings who would act freely but always go right. Clearly
, his failure to avail himself of this possibility is inconsistent with his bein
g both omnipotent and perfectly good.
According to Plantinga, Mackie is correct in thinking that there is nothing impo
ssible about a world in which people always freely choose to do right. That’s W4.
He is also correct in thinking that God’s only options were not “making innocent aut
omata and making beings who, in acting freely, would sometimes go wrong.” In other
words, worlds like W1 and W2 are not the only logically possible worlds. But Pl
antinga thinks he is mistaken in thinking that W3 is possible and in not recogni
zing important differences between W3 and W4. People can freely choose to do wha
t is right only when their actions are not causally determined.
We might wonder why God would choose to risk populating his new creation with fr
ee creatures if he knew there was a chance that human immorality could foul the
whole thing up. C. S. Lewis (1943, p. 52) offers the following answer to this qu
estion:
Why, then, did God give them free will? Because free will, though it makes evil
possible, is also the only thing that makes possible any love or goodness or joy
worth having. A world of automata—of creatures that worked like machines—would hard
ly be worth creating. The happiness which God designs for His higher creatures i
s the happiness of being freely, voluntarily united to Him and to each other…. And
for that they must be free. Of course, God knew what would happen if they used
their freedom the wrong way: apparently He thought it worth the risk.
Plantinga concurs. He writes,
A world containing creatures who are sometimes significantly free (and freely pe
rform more good than evil actions) is more valuable, all else being equal, than
a world containing no free creatures at all. Now God can create free creatures,
but he cannot cause or determine them to do only what is right. For if he does s
o, then they are not significantly free after all; they do not do what is right
freely. To create creatures capable of moral good, therefore, he must create cre
atures capable of moral evil; and he cannot leave these creatures free to perfor
m evil and at the same time prevent them from doing so…. The fact that these free
creatures sometimes go wrong, however, counts neither against God’s omnipotence no
r against his goodness; for he could have forestalled the occurrence of moral ev
il only by excising the possibility of moral good. (Plantinga 1974, pp. 166-167)
According to his Free Will Defense, God could not eliminate the possibility of m
oral evil without at the same time eliminating some greater good.
5. Divine Omnipotence and the Free Will Defense
Some scholars maintain that Plantinga has rejected the idea of an omnipotent God
because he claims there are some things God cannot do—namely, logically impossibl
e things. Plantinga, however, doesn’t take God’s omnipotence to include the power to
do the logically impossible. He reasons as follows. Can God create a round squa
re? Can he make 2 + 2 = 5? Can he create a stick that is not as long as itself?
Can he make contradictory statements true? Can he make a rock so big he can’t lift
it? In response to each of these questions, Plantinga’s answer is “No.” Each of the s
cenarios depicted in these questions is impossible: the objects or events in que
stion couldn’t possibly exist. Omnipotence, according to Plantinga, is the power t
o do anything that is logically possible. The fact that God cannot do the logica
lly impossible is not, Plantinga claims, a genuine limitation of God’s power. He w
ould urge those uncomfortable with the idea of limitations on God’s power to think
carefully about the absurd implications of a God who can do the logically impos
sible. If you think God really can make a round square, Plantinga would like to
know what such a shape would look like. If God can make 2 + 2 = 5, then what wou
ld 2 + 3 equal? If God can make a rock so big that he can’t lift it, exactly how b
ig would that rock be? What Plantinga would really like to see is a stick that i
s not as long as itself. Each of these things seems to be absolutely, positively
impossible.
Many theists maintain that it is a mistake to think that God’s omnipotence require
s that the blank in the following sentence must never be filled in:
(23) God is not able to ______________.
According to orthodox theism, all of the following statements (and many more lik
e them) are true.
(24) God is not able to lie.
(25) God is not able to cheat.
(26) God is not able to steal.
(27) God is not able to be unjust.
(28) God is not able to be envious.
(29) God is not able to fail to know what is right.
(30) God is not able to fail to do what he knows to be right.
(31) God is not able to have false beliefs about anything.
(32) God is not able to be ignorant.
(33) God is not able to be unwise.
(34) God is not able to cease to exist.
(35) God is not able to make a mistake of any kind.
According to classical theism, the fact that God cannot do any of these things i
s not a sign of weakness. On the contrary, theists claim, it is an indication of
his supremacy and uniqueness. These facts reveal that God is, in St. Anselm’s (10
33-1109 A.D.) words, “that being than which none greater can be conceived.” Planting
a adds the following two items to the list of things God cannot do.
(36) God is not able to contradict himself.
(37) God is not able to make significantly free creatures and to causally determ
ine that they will always choose what is right and avoid what is wrong.
These inabilities follow not from God’s omnipotence alone but from his omnipotence
in combination with his omniscience, moral perfection and the other divine perf
ections God possesses.
6. An Objection: Free Will and Natural Evil
At this point, someone might raise the following objection.
Plantinga can’t put all the blame for pain and suffering on human beings. Although
much of the evil in this world results from the free choices people make, some
of it does not. Cancer, AIDS, famines, earthquakes, tornadoes, and many other ki
nds of diseases and natural disasters are things that happen without anybody cho
osing to bring them about. Plantinga’s Free Will Defense, then, cannot serve as a
morally sufficient reason for God’s allowing disease and natural disasters.
This objection leads us to draw a distinction between the following two kinds of
evil and suffering:
(38) Moral evil =df evil or suffering that results from the immoral choices of f
ree creatures.
(39) Natural evil =df evil or suffering that results from the operations of natu
re or nature gone awry.
According to Edward Madden and Peter Hare (1968, p. 6), natural evil includes
the terrible pain, suffering, and untimely death caused by events like fire, flo
od, landslide, hurricane, earthquake, tidal wave, and famine and by diseases lik
e cancer, leprosy and tetanus—as well as crippling defects and deformities like bl
indness, deafness, dumbness, shriveled limbs, and insanity by which so many sent
ient beings are cheated of the full benefits of life.
Moral evil, they continue, includes
both moral wrong-doing such as lying, cheating, stealing, torturing, and murderi
ng and character defects like greed, deceit, cruelty, wantonness, cowardice, and
selfishness. (ibid.)
It seems that, although Plantinga’s Free Will Defense may be able to explain why G
od allows moral evil to occur, it cannot explain why he allows natural evil. If
God is going to allow people to be free, it seems plausible to claim that they n
eed to have the capacity to commit crimes and to be immoral. However, it is not
clear that human freedom requires the existence of natural evils like deadly vir
uses and natural disasters. How would my free will be compromised if tomorrow Go
d completely eliminated cancer from the face of the Earth? Do people really need
to die from heart disease and flash floods in order for us to have morally sign
ificant free will? It is difficult to see that they do. So, the objection goes,
even if Plantinga’s Free Will Defense explains why God allows moral evil, it does
not explain why he allows natural evil.
Plantinga, however, thinks that his Free Will Defense can be used to solve the l
ogical problem of evil as it pertains to natural evil. Here is a possible reason
God might have for allowing natural evil:
(MSR2) God allowed natural evil to enter the world as part of Adam and Eve’s punis
hment for their sin in the Garden of Eden.
(Those familiar with Plantinga’s work will notice that this is not the same reason
Plantinga offers for God’s allowing natural evil. They will also be able to guess
why a different reason was chosen in this article.) The sin of Adam and Eve was
a moral evil. (MSR2) claims that all natural evil followed as the result of the
world’s first moral evil. So, if it is plausible to think that Plantinga’s Free Wil
l Defense solves the logical problem of evil as it pertains to moral evil, the c
urrent suggestion is that it is plausible also to think that it solves the logic
al problem of evil as it pertains to natural evil because all of the worlds evil
s have their source in moral evil.
(MSR2) represents a common Jewish and Christian response to the challenge posed
by natural evil. Death, disease, pain and even the tiresome labor involved in gl
eaning food from the soil came into the world as a direct result of Adam and Eve’s
sin. The emotional pain of separation, shame and broken relationships are also
consequences that first instance of moral evil. In fact, according to the first
chapter of Genesis, animals in the Garden of Eden didn’t even kill each other for
food before the Fall. In the description of the sixth day of creation God says t
o Adam and Eve,
I give you every seed-bearing plant on the face of the whole earth and every tre
e that has fruit with seed in it. They will be yours for food. And to all the be
asts of the earth and all the birds of the air and all the creatures that move o
n the ground—everything that has the breath of life in it—I give every green plant f
or food. (Gen. 1:29-30, NIV)
In other words, the Garden of Eden is pictured as a peaceful, vegetarian commune
until moral evil entered the world and brought natural evil with it. It seems,
then, that the Free Will Defense might be adapted to rebut the logical problem o
f natural evil after all.
Some might think that (MSR2) is simply too far-fetched to be taken seriously. [I
f you think (MSR2) is far-fetched, see Plantinga s (1974, pp. 191-193) own sugge
stions about who is responsible for natural evil.] Natural disasters, it will be
said, bear no essential connection to human wrongdoing, so it is absurd to thin
k that moral evil could somehow bring natural evil into the world. Moreover, (MS
R2) would have us believe that there were real persons named Adam and Eve and th
at they actually performed the misdeeds attributed to them in the book of Genesi
s. (MSR2) seems to be asking us to believe things that only a certain kind of th
eist would believe. The implausibility of (MSR2) is taken by some to be a seriou
s defect.
7. Evaluating the Free Will Defense
What should we make of Plantinga’s Free Will Defense? Does it succeed in solving t
he logical problem of evil as it pertains to either moral or natural evil? In or
der to answer these questions, let’s briefly consider what it would take for any r
esponse to the logical problem of evil to be successful. Recall that the logical
problem of evil can be summarized as the following claim:
(16) It is not possible for God and evil to co-exist.
When someone claims
(40) Situation x is impossible,
what is the least that you would have to prove in order to show that (40) is fal
se? If you could point to an actual instance of the type of situation in questio
n, that would certainly prove that (40) is false. But you don’t even need to troub
le yourself with finding an actual x. All you need is a possible x. The claim
(41) Situation x is possible
is the contradictory of (40). The two claims are logical opposites. If one is tr
ue, the other is false; if one is false, the other is true. If you can show that
x is merely possible, you will have refuted (40).
How would you go about finding a logically possible x? Philosophers claim that y
ou only need to use your imagination. If you can conceive of a state of affairs
without there being anything contradictory about what you’re imagining, then that
state of affairs must be possible. In a word, conceivability is your guide to po
ssibility.
Since the logical problem of evil claims that it is logically impossible for God
and evil to co-exist, all that Plantinga (or any other theist) needs to do to c
ombat this claim is to describe a possible situation in which God and evil co-ex
ist. That situation doesn’t need to be actual or even realistic. Plantinga doesn’t n
eed to have a single shred of evidence supporting the truth of his suggestion. A
ll he needs to do is give a logically consistent description of a way that God a
nd evil can co-exist. Plantinga claims God and evil could co-exist if God had a
morally sufficient reason for allowing evil. He suggests that God’s morally suffic
ient reason might have something to do with humans being granted morally signifi
cant free will and with the greater goods this freedom makes possible. All that
Plantinga needs to claim on behalf of (MSR1) and (MSR2) is that they are logical
ly possible (that is, not contradictory).
Does Plantinga’s Free Will Defense succeed in describing a possible state of affai
rs in which God has a morally sufficient reason for allowing evil? It certainly
seems so. In fact, it appears that even the most hardened atheist must admit tha
t (MSR1) and (MSR2) are possible reasons God might have for allowing moral and n
atural evil. They may not represent God’s actual reasons, but for the purpose of b
locking the logical problem of evil, it is not necessary that Plantinga discover
God’s actual reasons. In the last section we noted that many people will find (MS
R2)’s explanation of natural evil extremely difficult to believe because it assume
s the literal existence of Adam and Eve and the literal occurrence of the Fall.
However, since (MSR2) deals with the logical problem of evil as it pertains to n
atural evil (which claims that it is logically impossible for God and natural ev
il to co-exist), it only needs to sketch a possible way for God and natural evil
to co-exist. The fact that (MSR2) may be implausible does not keep it from bein
g possible. Since the situation described by (MSR2) is clearly possible, it appe
ars that it successfully rebuts the logical problem of evil as it pertains to na
tural evil.
Since (MSR1) and (MSR2) together seem to show contra the claims of the logical p
roblem of evil how it is possible for God and (moral and natural) evil to co-exi
st, it seems that the Free Will Defense successfully defeats the logical problem
of evil.
8. Was Plantinga’s Victory Too Easy?
Some philosophers feel that Plantinga’s apparent victory over the logical problem
of evil was somehow too easy. His solution to the logical problem of evil leaves
them feeling unsatisfied and suspicious that they have been taken in by some ki
nd of sleight of hand. For example, J. L. Mackie one of the most prominent athei
st philosophers of the mid-twentieth-century and a key exponent of the logical p
roblem of evil has this to say about Plantinga’s Free Will Defense:
Since this defense is formally [that is, logically] possible, and its principle
involves no real abandonment of our ordinary view of the opposition between good
and evil, we can concede that the problem of evil does not, after all, show tha
t the central doctrines of theism are logically inconsistent with one another. B
ut whether this offers a real solution of the problem is another question. (Mack
ie 1982, p. 154)
Mackie admits that Plantinga’s defense shows how God and evil can co-exist, that i
s, it shows that “the central doctrines of theism” are logically consistent after al
l. However, Mackie is reluctant to attribute much significance to Plantinga’s acco
mplishment. He expresses doubt about whether Plantinga has adequately dealt with
the problem of evil.
Part of Mackie’s dissatisfaction probably stems from the fact that Plantinga only
gives a possible reason for why God might have for allowing evil and suffering a
nd does not provide any evidence for his claims or in any way try to make them p
lausible. Although sketching out mere possibilities without giving them any evid
ential support is typically an unsatisfactory thing to do in philosophy, it is n
ot clear that Mackie’s unhappiness with Plantinga is completely warranted. It was,
after all, Mackie himself who characterized the problem of evil as one of logic
al inconsistency:
Here it can be shown, not that religious beliefs lack rational support, but that
they are positively irrational, that several parts of the essential theological
doctrine are inconsistent with one another. (Mackie 1955, p. 200)
In response to this formulation of the problem of evil, Plantinga showed that th
is charge of inconsistency was mistaken. Even Mackie admits that Plantinga solve
d the problem of evil, if that problem is understood as one of inconsistency. It
is, therefore, difficult to see why Plantinga’s Free Will Defense should be found
wanting if that defense is seen as a response to the logical problem of evil. A
s an attempt to rebut the logical problem of evil, it is strikingly successful.
The dissatisfaction many have felt with Plantinga’s solution may stem from a desir
e to see Plantinga’s Free Will Defense respond more generally to the problem of ev
il and not merely to a single formulation of the problem. As an all-around respo
nse to the problem of evil, the Free Will Defense does not offer us much in the
way of explanation. It leaves several of the most important questions about God
and evil unanswered. The desire to see a theistic response to the problem of evi
l go beyond merely undermining a particular atheological argument is understanda
ble. However, we should keep in mind that all parties admit that Plantinga’s Free
Will Defense successfully rebuts the logical problem of evil as it was formulate
d by atheists during the mid-twentieth-century.
If there is any blame that needs to go around, it may be that some of it should
go to Mackie and other atheologians for claiming that the problem of evil was a
problem of inconsistency. The ease with which Plantinga undermined that formulat
ion of the problem suggests that the logical formulation did not adequately capt
ure the difficult and perplexing issue concerning God and evil that has been so
hotly debated by philosophers and theologians. In fact, this is precisely the me
ssage that many philosophers took away from the debate between Plantinga and the
defenders of the logical problem of evil. They reasoned that there must be more
to the problem of evil than what is captured in the logical formulation of the
problem. It is now widely agreed that this intuition is correct. Current discuss
ions of the problem focus on what is called “the probabilistic problem of evil” or “th
e evidential problem of evil.” According to this formulation of the problem, the e
vil and suffering (or, in some cases, the amounts, kinds and distributions of ev
il and suffering) that we find in the world count as evidence against the existe
nce of God (or make it improbable that God exists). Responding to this formulati
on of the problem requires much more than simply describing a logically possible
scenario in which God and evil co-exist.
9. Other Responses to the Logical Problem of Evil
Plantinga’s Free Will Defense has been the most famous theistic response to the lo
gical problem of evil because he did more to clarify the issues surrounding the
logical problem than anyone else. It has not, however, been the only such respon
se. Other solutions to the problem include John Hick’s (1977) soul-making theodicy
. Hick rejects the traditional view of the Fall, which pictures humans as being
created in a finitely perfect and finished state from which they disastrously fe
ll away. Instead, Hick claims that human beings are unfinished and in the midst
of being made all that God intended them to be. The long evolutionary process ma
de humans into a distinguishable species capable of reasoning and responsibility
, but they must now (as individuals) go through a second process of “spiritualizat
ion” or “soul-making,” during which they become “children of God.” According to Hick, the
suffering and travails of this life are part of the divine plan of soul-making.
A world full of suffering, trials and temptations is more conducive to the proce
ss of soul-making than a world full of constant pleasure and the complete absenc
e of pain. Hick (1977, pp. 255-256) writes,
The value-judgment that is implicitly being invoked here is that one who has att
ained to goodness by meeting and eventually mastering temptations, and thus by r
ightly making responsible choices in concrete situations, is good in a richer an
d more valuable sense than would be one created ab initio in a state either of i
nnocence or of virtue…. I suggest, then, that it is an ethically reasonable judgme
nt… that human goodness slowly built up through personal histories of moral effort
has a value in the eyes of the Creator which justifies even the long travail of
the soul-making process.
Unlike Plantinga’s response to the logical problem of evil, which is merely a “defen
se” (that is, a negative attempt to undermine a certain atheological argument with
out offering a positive account of why God allows evil and suffering), Hick’s resp
onse is a “theodicy” (that is, a more comprehensive attempt to account for why God i
s justified in allowing evil and suffering).
Eleonore Stump (1985) offers another response to the problem of evil that brings
a range of distinctively Christian theological commitments to bear on the issue
. She claims that a world full of evil and suffering is “conducive to bringing abo
ut both the initial human [receipt of God s gift of salvation] and also the subs
equent process of sanctification” (Stump 1985, p. 409). She writes,
Natural evil—the pain of disease, the intermittent and unpredictable destruction o
f natural disasters, the decay of old age, the imminence of death—takes away a per
son’s satisfaction with himself. It tends to humble him, show him his frailty, mak
e him reflect on the transience of temporal goods, and turn his affections towar
ds other-worldly things, away from the things of this world. No amount of moral
or natural evil, of course, can guarantee that a man will [place his faith in Go
d]…. But evil of this sort is the best hope, I think, and maybe the only effective
means, for bringing men to such a state. (Stump 1985, p. 409)
Stump claims that, although the sin of Adam—and not any act of God—first brought mor
al and natural evil into this world, God providentially uses both kinds of evil
in order to bring about the greatest good that a fallen, sinful human being can
experience: a repaired will and eternal union with God.
The responses of both Hick and Stump are intended to cover not only the logical
problem of evil but also any other formulation of the problem as well. Thus, som
e of those dissatisfied with Plantinga’s merely defensive response to the problem
of evil may find these more constructive, alternative responses more attractive.
Regardless of the details of these alternatives, the fact remains that all they
need to do in order to rebut the logical problem of evil is to describe a logic
ally possible way that God and evil can co-exist. A variety of morally sufficien
t reasons can be proposed as possible explanations of why a perfect God might al
low evil and suffering to exist. Because the suggestions of Hick and Stump are c
learly logically possible, they, too, succeed in undermining the logical problem
of evil.
10. Problems with the Free Will Defense
A. Even though it is widely agreed that Plantinga’s Free Will Defense describes a
state of affairs that is logically possible, some of the details of his defense
seem to conflict with important theistic doctrines. One point of conflict concer
ns the possibility of human free will in heaven. Plantinga claims that if someon
e is incapable of doing evil, that person cannot have morally significant free w
ill. He also maintains that part of what makes us the creatures we are is that w
e possess morally significant freedom. If that freedom were to be taken away, we
might very well cease to be the creatures we are. However, consider the sort of
freedom enjoyed by the redeemed in heaven. According to classical theism, belie
vers in heaven will somehow be changed so that they will no longer commit any si
ns. It is not that they will contingently always do what is right and contingent
ly always avoid what is wrong. They will somehow no longer be capable of doing w
rong. In other words, their good behavior will be necessary rather than continge
nt.
This orthodox view of heaven poses the following significant challenges to Plant
inga’s view:
(i) If heavenly dwellers do not possess morally significant free will and yet th
eir existence is something of tremendous value, it is not clear that God was jus
tified in creating persons here on Earth with the capacity for rape, murder, tor
ture, sexual molestation, and nuclear war. It seems that God could have actualiz
ed whatever greater goods are made possible by the existence of persons without
allowing horrible instances of evil and suffering to exist in this world.
(ii) If possessing morally significant free will is essential to human nature, i
t is not clear how the redeemed can lose their morally significant freedom when
they get to heaven and still be the same people they were before.
(iii) If despite initial appearances heavenly dwellers do possess morally signif
icant free will, then it seems that it is not impossible for God to create genui
nely free creatures who always (of necessity) do what is right.
In other words, it appears that W3 isn’t impossible after all. If W3 is possible,
an important plank in Plantinga’s Free Will Defense is removed. None of these chal
lenges undermines the basic point established above that Plantinga’s Free Will Def
ense successfully rebuts the logical problem of evil. However, they reveal that
some of the central claims of his defense conflict with other important theistic
doctrines. Although Plantinga claimed that his Free Will Defense offered merely
possible and not necessarily actual reasons God might have for allowing evil an
d suffering, it may be difficult for other theists to embrace his defense if it
runs contrary to what theism says is actually the case in heaven.
B. Another problem facing Plantinga’s Free Will Defense concerns the question of G
od’s free will. God, it seems, is incapable of doing anything wrong. Thus, it does
not appear that, with respect to any choice of morally good and morally bad opt
ions, God is free to choose a bad option. He seems constitutionally incapable of
choosing (or even wanting) to do what is wrong. According to Plantinga’s descript
ion of morally significant free will, it does not seem that God would be signifi
cantly free. Plantinga suggests that morally significant freedom is necessary in
order for one’s actions to be assessed as being morally good or bad. But then it
seems that God’s actions could not carry any moral significance. They could never
be praiseworthy. That certainly runs contrary to central doctrines of theism.
If, as theists must surely maintain, God does possess morally significant freedo
m, then perhaps this sort of freedom does not preclude an inability to choose wh
at is wrong. But if it is possible for God to possess morally significant freedo
m and for him to be unable to do wrong, then W3 once again appears to be possibl
e after all. Originally, Plantinga claimed that W3 is not a logically possible w
orld because the description of that world is logically inconsistent. If W3 is p
ossible, then the complaint lodged by Flew and Mackie above that God could (and
therefore should) have created a world full of creatures who always did what is
right is not answered.
There may be ways for Plantinga to resolve the difficulties sketched above, so t
hat the Free Will Defense can be shown to be compatible with theistic doctrines
about heaven and divine freedom. As it stands, however, some important challenge
s to the Free Will Defense remain unanswered. It is also important to note that,
simply because Plantinga’s particular use of free will in fashioning a response t
o the problem of evil runs into certain difficulties, that does not mean that ot
her theistic uses of free will in distinct kinds of defenses or theodicies would
face the same difficulties.