172

P A R T II T R A D I T I O N A L A R G U M E N T S F O R T H E E X I S T E N C E O F G O D

objects do not stand in causal relations. Therefore,

the transcendent cause of the origin of the universe
must be of the order of mind. Third, this same
conclusion is also implied by the fact that we have
in this case the origin of a temporal effect from a
timeless cause. If the cause of the origin of the universe were an impersonal set of necessary and sufficient conditions, it would be impossible for the
cause to exist without its effect. For if the necessary
and sufficient conditions of the effect are timelessly

given, then their effect must be given as well. The

only way for the cause to be timeless and changeless
but for its effect to originate anew a finite time ago
is for the cause to be a personal agent who freely
chooses to bring about an effect without antecedent
determining conditions. Thus we are brought, not
merely to a transcendent cause of the universe, but
to its Personal Creator. He is, as Leibniz maintained, the Sufficient Reason why anything exists
rather than nothing.

NOTES
1. This should not be taken to mean that the density
of the universe takes on a value of H0 but rather
that the density of the universe is expressed by a
ratio of mass to volume in which the volume is
zero; since division by zero is impermissible, the
density is said to be infinite in this sense.
2. Quentin Smith, The Uncaused Beginning of the
Universe, in Theism, Atheism and Big Bang
Cosmology, by William Lane Craig and Quentin
Smith (Oxford: Clarendon, 1993), p. 120.

3. Arthur Eddington, The Expanding Universe

(New York: Macmillan, 1933), p. 124.
4. Ibid., p. 178.
5. Stephen Hawking and Roger Penrose, The Nature
of Space and Time, The Isaac Newton Institute
Series of Lectures (Princeton, N.J.: Princeton
University Press, 1996), p. 20.
6. Richard Schlegel, Time and Thermodynamics,
in The Voices of Time, ed. J. T. Fraser (London:
Penguin, 1948), p. 511.

II.B.5
A Critique of the Kalam Cosmological Argument
PAUL DRAPER

Paul Draper is professor of philosophy at Purdue University and the author of several important essays in
the philosophy of religion. In this article he analyzes William Lane Craigs philosophical defense of the
kalam cosmological argument. Draper contends that Craigs defense fails, both because it fails to establish
that the universe had a beginning and because it rests on an equivocation of the phrase begins to exist.

Copyright Paul Draper 1997. Used by permission of the author.

PAUL DRAPER A CRITIQUE OF THE KAL AM COSMOLOGICAL ARGUMENT

Epistemology begins in doubt, ethics in conflict, and

metaphysics in wonder.
In a recent book,1 William Lane Craig offers a philosophical and scientific defense of a very old and
very wonderful argument: the kalam cosmological
argument. Unlike other cosmological arguments,
the kalam argument bases its conclusion that the
universe has a cause of its existence on the premise
that the universe began to exist a finite time ago.
Craig calls it the kalam cosmological argument
because kalam is the name of a theological movement within Islam that used reason, including this
argument, to defend the Muslim faith against philosophical objections. After being fully developed by
Arab thinkers like al-Kindi and al-Ghazali, the
argument eventually made its way to the West,
where it was rejected by St. Thomas Aquinas and
defended by St. Bonaventure.2 My focus in this
paper will be on Craigs philosophical defense of
the argument. I will try to show that this defense
fails, both because it fails to establish that the universe had a beginning and because it commits the
fallacy of equivocation.
Compare the following two cosmological
arguments, each of which concludes that the universe has a cause of its existence:
1. Every contingent thing (including things that
are infinitely old) has a cause of its existence.
2. The universe is contingent.
3. Therefore, the universe has a cause of its
existence.
1. Everything that begins to exist has a cause of its
existence.
2. The universe began to exist.
3. Therefore, the universe has a cause of its
existence.
The first of these arguments is sometimes called the
argument from contingency. It was suggested by
Aristotle, clearly formulated by Arabic philosophers like ibn Sina, and later championed in the
West by St. Thomas Aquinas. I find it completely
unpersuasive. For although the second premise is
clearly true (so long as contingent means logically

173

contingent), I do not find the first premise appealing at all. If something is infinitely old, then it has
always existed, and its hard to see why something
that has always existed requires a cause of its existence, even if it is logically possible that it not have
existed. (Indeed, its not even clear that something
that has always existed could have a cause of its
existence.)
The second of these arguments is the kalam
cosmological argument. This argument avoids
the weakness of the argument from contingency by
denying that the universe is infinitely old and maintaining that the universe needs a cause, not because it
is contingent, but rather because it had a beginning.
In other words, it replaces the weak premise that
every contingent thing needs a cause of its existence
with the compelling premise that everything that
begins to exist needs a cause of its existence. Of
course, a price must be paid for strengthening the
first premise: the second premisethat the universe
began to existis not by a long shot as unquestionably true as the claim that the universe is contingent.
Craig, however, provides a spirited and plausible defense of this premise. He offers four
arguments in support of it, two of which are philosophical (armchair cosmology at its best) and two
of which are scientific (but still interesting). Both
philosophical arguments depend on a distinction
between a potential infinite and an actual infinite.
A potential infinite is a series or collection that can
increase forever without limit but is always finite
(e.g., the set of events that have occurred since
the birth of my daughter or the set of completed
years after 1000 BCE). An actual infinite is a set
of distinct things (real or not) whose number is
actually infinite (e.g., the set of natural numbers).
The first philosophical argument claims that there
cant be an infinite regress of events, because actual
infinites cannot exist in reality. According to the
second argument, an infinite regress of events is
impossible because, even if actual infinites could
exist in reality, they could not be formed by successive addition.
The first scientific argument is based on the evidence for the Big Bang theory, which seems to many
scientists to support the view that the universe had a

174

P A R T II T R A D I T I O N A L A R G U M E N T S F O R T H E E X I S T E N C E O F G O D

beginning. The second scientific argument appeals

to the Second Law of Thermodynamics. According
to this law, the amount of energy available to do
mechanical work always decreases in a closed system.
Thus, since the universe as a whole is a closed system
with a finite amount of such energy, an infinitely old
universe is incompatible with the fact that we have
not yet run out of such energythe universe has not
yet reached its equilibrium end state. Since Im no
scientist, I will focus my attention on Craigs philosophical arguments, beginning with the second one.
As Craig himself points out, his second philosophical argument is very similar to the argument
that Immanuel Kant uses to support the thesis of his
first antinomy:
If we assume that the world has no
beginning in time, then up to every given
moment an eternity has elapsed and there
has passed away in the world an infinite
series of successive states of things. Now
the infinity of a series consists in the fact
that it can never be completed through
successive synthesis. It thus follows that it is
impossible for an infinite world-series to
have passed away, and that a beginning
of the world is therefore a necessary
condition of the worlds existence.3
Craig formulates the argument as follows:
(i) The temporal series of events is a collection
formed by successive addition.
(ii) A collection formed by successive addition
cannot be an actual infinite.
(iii) Thus, the temporal series of events cannot be
an actual infinite. (from i and ii)
(iv) Therefore, the temporal regress of events is
finite. (from iii)4
This argument is closely related to Zenos paradoxes,
which depend on the claim that one cannot complete an infinite series of tasks one at a time since that
would imply an infinitieth member of the series. As
it stands, the argument is unconvincing. For while it
is true that one cannot start with a finite collection
and then by adding one new member at a time turn
it into an infinite collection (no matter how much

time one has available), nothing of the sort is

required in order for the past to be infinite. For if
the temporal regress of events is infinite, then the
universe has never had a finite number of past events.
Rather, it has always been the case that the collection of past events is infinite. Thus, if the temporal
regress of events is infinite, then the temporal series
of events is not an infinite collection formed by successively adding to a finite collection. Rather, it is a
collection formed by successively adding to an infinite collection. And surely it is not impossible to
form an infinite collection by successively adding to
an already infinite collection.
One might object that, if the temporal regress
of events is infinite, then there must be some event
E separated from the birth of my daughter by an
infinite number of intermediate events, in which
case the collection containing E and all those intermediate events would have to be an actually infinite
collection formed by successively adding to a finite
collection of events, namely the collection containing E as its only member. This objection fails
because it is simply not true that, if the temporal
regress of events is infinite, then there must be two
events separated by an infinite number of intermediate events. For consider the set of natural numbers. It is actually infinite, yet every member of it is
such that there is a finite number of members
between it and its first member.5
Craigs first philosophical argument is, I believe,
much more promising than his second. It bases its
conclusion that the temporal regress of physical
events must be finitethere must have been a first
physical eventon the premises that an actual infinite
cannot exist in reality and an infinite temporal regress
of events is an actual infinite.6 From this and the further claim that a first physical event could not have
been preceded by an eternal absolutely quiescent
physical universe, the conclusion is drawn that the
physical universe had a beginning. The first stage of
this argument can be formulated as follows:
a. No set of real things is actually infinite.
b. If there was no first event, then the set of all
real events occurring prior to the birth of my
daughter is actually infinite.
c. Therefore, there was a first event.

PAUL DRAPER A CRITIQUE OF THE KAL AM COSMOLOGICAL ARGUMENT

Craig defends premise (a) of this argument

by pointing out that the assumption that a set of
real things is actually infinite has paradoxical implications.7 For example, it implies that we could have a
library consisting of infinitely many black books
(each might be assigned an even number). We
could then add infinitely many red books (each
might be assigned an odd number) and yet not
increase the number of books in the library by a
single volume. Indeed, we could add infinitely
many different colors of books with infinitely many
books of each color (the red books could be assigned
rational numbers between 0 and 1, the black books
rational numbers between 1 and 2, and so on) and
not increase our collection by a single volume.
These paradoxes arise because the following
three statements constitute an inconsistent triad:
S1. A set has more members than any of its proper
subsets.
S2. If the members of two sets can be placed in
one-to-one correspondence, then neither set
has more members than the other.
S3. There are actually infinite sets.
For example, since the set of even numbers has
one-to-one correspondence with the set of natural
numbers and even with the set of rational numbers,
S2 implies that one could add infinitely many
red books or infinitely many books of each of
infinitely many different colors to the library without increasing the size of that librarys collection.
(One need only make sure that the additions are
denumerably infinite.) But of course S1 implies that
any such addition would increase the size of the
collection since the set of even numbers is a proper
subset both of the set of natural numbers and of
the set of rational numbers. Thus, two intuitively
appealing principles together imply a contradiction
on the assumption that there can be an actually
infinite collection of books. One way to avoid
this contradiction is to reject the assumption that
there can be an actually infinite collection of
books. So the underlying argument in defense of
the claim that no collection of real things is actually
infinite is simply that, since S1 and S2 are both true
of collections of real things, it follows that S3 is not

175

true of such collectionsno collections of real

things are actually infinite.
Craig claims that Georg Cantors theory of
transfinite numbers is consistent because it rejects
the first member of the triad. But this member is
not rejected because it can be proven false about
actually infinite sets, nor is the second member
accepted because it can be proven that if a oneto-one correspondence between the elements of
two actually infinite sets can be established then
the sets are equivalent. Rather, equivalent sets are
simply defined as sets having one-to-one
correspondence. Thus, while Cantors theory is a
consistent mathematical system, there is, according
to Craig, no reason to think that it has any interesting ontological implications. In particular, it does
not provide any reason to think that S1 is false
about actually infinite sets and hence provides no
justification for thinking that actual infinites can
exist in reality.8
Notice that, if Craig is right that past events are
real but future events are not, then his argument for
a first event does not commit him to the position
that there is a last event. For consider the following
parallel argument for the conclusion that there will
be a last event:
(a) No set of real things is actually infinite.
(b) If there will be no last event, then the set of all
real events occurring after the birth of my
daughter is actually infinite.
(c) Therefore, there will be a last event.
Since future events are not real, the second premise
of this argument is false. If there is no last event,
then the set of all real events occurring after
the birth of my daughter is merely potentially
infinitenot actually infinite. This collection can
increase in size indefinitely, but it will always be
finite. Past events, on the other hand, are all real.
So if there is no first past event, then the set of all
real past events is actually infinite, not potentially
infinite. Craig concludes that, although there may
be no last event, there must be a first event, and
hence, since matter cannot exist without events
occurring, it follows that the universe has not
always existedit began to exist.

176

P A R T II T R A D I T I O N A L A R G U M E N T S F O R T H E E X I S T E N C E O F G O D

Although this fascinating argument for the second premise of the kalam argument may be sound,
Craig has not given us adequate reason to believe it
is. The problem concerns the inconsistent triad
mentioned above. What Craig needs to do is to
show that, when it comes to collections of real
things, we should reject the third member of the
triad instead of S1 or S2. But he has not shown this.
S1 and S2 are certainly true for finite collections.
But its far from clear that they are true for all collections. Allow me to explain why.
Consider S1, which says that a set has more
members than any of its proper subsets. If more
means a greater number, then the claim that S1 is
true for actually infinite sets requires us to make
sense of claiming that actually infinite sets have a
number of members. But an actually infinite set
doesnt have a natural number of members or a
rational number of members or a real number of
members, so one such set cant have a greater natural or rational or real number of members than
another. Of course, an actually infinite set does
have a transfinite number of members. But transfinite numbers are what Cantor defines them to be.
And given his definition, it simply isnt true that
actually infinite sets have a greater transfinite number of members than all of their proper subsets. We
could say that an actually infinite set has a greater
infinite number of members than all of its proper
subsets, but Craig gives us no theory of infinite
numbers that would justify that claim.
Of course, Craig might claim that no such theory
is necessary, that we dont even need to make use of
the word number here; for its just obvious that, in
some sense of the word more, any set that has every
member that another set has and some members it
doesnt have has more members than the other set. I
agree this is obvious, but in the case of infinite sets, this
is obvious only because more can just mean has
every member the other set has and some members it
doesnt have. If, however, we grant Craig that S1 is
true on these grounds, then why accept S2? Why not
claim instead that actually infinite collections of real
objects are possible, but the fact that two of them have
one-to-one correspondence is not a good reason to
believe that neither has more members than the
other? Why, for example, is it more reasonable to

believe that actually infinite libraries are impossible

than to believe that, although they are possible, one
such library can have more books than a second
despite the fact that the books in the first can be placed
in one-to-one correspondence with the books in the
second? Craig provides no good answer to these questions. Obviously he cannot all of a sudden appeal to
Cantors theory to justify accepting S2. For that
would commit him to rejecting S1. And since,
when infinite sets are compared, the word more cannot mean what it does when finite sets are compared,
the fact that S2 is true for finite sets is not by itself a
good reason to believe that it is true for all sets.
So Craig fails to show that S1 and S2 are both
true of all collections of real objects, and hence he
fails to show that actually infinite collections of real
objects are impossible. Therefore, his first philosophical argument, like his second, fails to establish
that an infinite regress of events is impossible and so
fails to establish that the universe began to exist.
This leaves us with Craigs scientific arguments.
Since I lack the expertise to evaluate these arguments, lets assume, for the sake of argument, that
they succeed and hence that the universe did begin
to exist. Must we then conclude that the kalam
argument succeeds? This would be a profound
result. Granted, this argument doesnt get all the
way to Gods existence. But accepting its conclusion does require rejecting naturalismsince nothing can be a cause of its own existence, a cause
outside the natural world would be required.
As wonderful as this conclusion is, I do not
believe that Craigs defense of the kalam argument
justifies accepting it, even assuming that his scientific
arguments are sound. This is because Craig commits
the fallacy of equivocation. The verb to begin has a
narrow or strict sense and a broad or loose sense. In
the narrow sense, to begin means to begin within
time. When used in this way, x begins to exist
implies that there was a time at which x did not
exist and then a later time at which x exists. But to
begin can also mean to begin either within or with
time. When used in this way, x begins to exist
does not imply that there was a time at which x did
not exist, because the past may itself be finite in which
case something that begins to exist at the first moment
in time is such that there never was a time at which it

PAUL DRAPER A CRITIQUE OF THE KAL AM COSMOLOGICAL ARGUMENT

did not existit begins with time rather than within

time. Now consider the two premises of the kalam
argument in the fight of this distinction.
The second premise is that the universe began to
exist. All of Craigs arguments in favor of this premise,
including his scientific ones, would be unsound if one
interpreted began to exist in the second premise as
meaning began to exist within time. For nothing in
these arguments counts against a relational view of
time. And on a relational view of time, a first temporal event is simultaneous with a first moment in time.
This would mean that, if the temporal series of past
events is finite, then the universe began to exist with
time. Indeed, if anything, the arguments in favor of
the second premise support a beginning with time.
For if an infinite regress of events is an actual infinite
and for that reason impossible, then it would seem
that an infinite past would be an actual infinite and
for that reason impossible. Moreover, one of Craigs
scientific arguments appeals to an interpretation of the
Big Bang Theory according to which time did not
exist before the big bang. So the most that Craig
has established is that the universe began to exist
either within or with time.
The first premise is that anything that begins to
exist has a cause of its existence. What does begins
to exist mean here? Craig defends this premise by
claiming that it is an empirical generalisation enjoying the strongest support experience affords.9 But
experience only supports the claim that anything that
begins to exist within time has a cause of its existence. For we have no experience whatsoever of
things beginning to exist with time.10 Such things
would require timeless causes. And even if it is conceptually possible for a temporal event to have a

177

timeless cause, we certainly have no experience of

this. Of course, Craig also claims that premise (1) is
intuitively obviousthat it needs no defense at all.
But it is far from obvious that a universe that begins
to exist with time needs a cause of its existence. Like
an infinitely old universe, a universe that begins to
exist with time has always existedfor any time t,
the universe existed at t. And once again, its far from
obvious that something that has always existed
requires a cause for its existence. Its not even clear
that such a thing could have a cause of its existence.
So in order to be justified in believing both of
the premises of the argumentjustified, that is,
solely on the basis of Craigs defense of those
premiseswe would need to equivocate on the
meaning of begins to exist. We would need to
use this term in the narrow sense in the first premise
and in the broad sense in the second premise. But
then the conclusion of the argument would not
follow from its premises. Thus, Craig commits the
fallacy of equivocation.11
Do my objections to Craigs defense of the kalam
argument prove that it is doomed? I dont think so.
The argument remains promising. Perhaps, for example, it could be shown that an absolute theory of time
is correct, and that such a theory, together with scientific or new philosophical evidence against an infinitely old universe, implies a beginning of the
universe within time. Or perhaps it could be shown
that the universe began to exist with time and that
even something that begins to exist with time
requires a cause of its existence. So my conclusion is
not that the kalam argument should be dismissed. It is
just that it has not yet been adequately defended. I
still wonder whether the argument is a good one.

NOTES
1. William Lane Craig, The Kalam Cosmological
Argument (New York: Harper & Row Publishers),
1979.
2. For a brief but interesting history of the argument,
see Craig, Part I.
3. Immanuel Kant, Critique of Pure Reason, trans.
Norman Kemp Smith (London: Macmillan & Co.,
1929), p. 396. Quoted by Craig on p. 189.

4. Craig, p. 103.
5. Cf. Quentin Smith, Infinity and the Past, in
Theism, Atheism, and Big Bang Cosmology, ed.
William Lane Craig and Quentin Smith (Oxford:
Clarendon Press, 1993), pp. 7883; Antony Flew,
The Case for God Challenged, in Does God
Exist?: The Great Debate, ed. J. P. Moreland and Kai
Nielsen (Nashville: Thomas Nelson Publishers,

178

6.
7.
8.
9.

10.

11.

P A R T II T R A D I T I O N A L A R G U M E N T S F O R T H E E X I S T E N C E O F G O D

1990), p. 164; and Keith Parsons, Is There a Case

for Christian Theism? in Does God Exist?: The
Great Debate, p. 187.
Craig, p. 69.
Craig, pp. 8287.
Craig, pp. 9495.
Craig, p. 145. Craig also suggests here that premise
(1) could be defended by appealing to an a priori
category of causality. Such Kantian maneuvering
does not seem very promising in this context. For
in order to reconcile it with the realism pre- argument, one would need
supposed by the kalam
to claim that the causal principle must, as a
necessary precondition of thought, hold without
exception in the noumenal world!
Cf. Quentin Smith, The Uncaused Beginning of
the Universe, in Theism, Atheism, and Big Bang
Cosmology, p. 123.
In The Caused Beginning of the Universe (in
Theism, Atheism, and Big Bang Cosmology) Craig
denies that his inference is equivocal on the

grounds that our conviction of the truth of the

causal principle is not based upon an inductive
survey of existents in space-time, but rather upon
the metaphysical intuition that something cannot
come out of nothing (p. 147). Of course, he did
appeal to such a survey in his book, but Craig
claims that this was just a last-ditch defence of the
principle designed to appeal to the hard-headed
empiricist who resists the metaphysical intuition
that properly grounds our conviction of the
principle (p. 147, note 13). This response to the
charge of equivocation is not at all convincing. For
metaphysical intuitions about contingent matters
are notoriously unreliablethats why so many
contemporary philosophers are, quite justifiably,
hard-headed empiricists. Further, at the risk of
committing the genetic fallacy, it is worth pointing
out that it is probably our experience of things
beginning to exist within time that causes some of
us to have the metaphysical intuition that something cannot come out of nothing.

II.C THE TELEOLOGICAL ARGUMENT FOR

THE EXISTENCE OF GOD
THE TELEOLOGICAL ARGUMENT for the existence of God begins with
the premise that the world exhibits intelligent purpose, order, or other marks of
design, and it proceeds to the conclusion that there must be or probably is a divine
intelligence, a supreme designer, to account for the observed or perceived intelligent purpose or order. Although core ideas of the argument can be found in
Plato, in the Bible (Rom. 1), and in Cicero, the most well-known treatment of
it is found in William Paleys Natural Theology (1802). In his opening chapter,
included here as our first selection, he offers his famous watch argument, which
begins as follows:
In crossing a heath, suppose I pitched my foot against a stone, and were
asked how the stone came to be there, I might possibly answer, that for
anything I knew to the contrary, it had lain there for ever; nor would it,
perhaps, be very easy to show the absurdity of this answer. But suppose I
found a watch upon the ground, and it should be inquired how the watch
happened to be in that place, I should hardly think of the answer which I
had before giventhat, for anything I knew, the watch might have always
been there. Yet why should not this answer serve for the watch as well as
for the stone? Why is it not as admissible in the second case, as in the first?