WHY TRADITIONAL COSMOLOGICAL ARGUMENTS DON’T WORK;
AND A SKETCH OF A NEW ONE THAT DOES
Bruce Reichenbach has done a masterful job of surveying
traditional versions of the cosmological argument, as well as attempting to
meet the standard objections to them. In
addition, he has defended his own version of the argument. I will
attempt to show that his argument, along with all other traditional
cosmological arguments, don’t work, and then go on to sketch a new one that
does.
I. Reichenbach’s “Deductive Cosmological Argument from
Contingency”
Reichenbach’s argument has as its first premise that
1. A contingent being (a being which, if it exists, can not-exist)
exists. (p. 9)
Because this premise can be known only via sense experience, it
keeps his argument from being
completely a priori. This is not a serious matter since no one but a
complete skeptic about the senses would doubt that there exists at least one
contingent being, such as a chair or an apple. In the course of defending this
argument, Reichenbach eventually, on p. 11, makes this existent contingent
being the aggregate of all existent contingent beings, that is, the universe.
The most controversial premise of the argument is
2. This contingent being
[i.e. the universe] has a cause or explanation of its existence.
Herein appeal is made to a
version of the principle of sufficient reason (hereafter PSR) or Principle of
Causation (hereafter PC), which holds that for every existent contingent being
there is a cause or explanation of its existence. Once this principle is
granted, the rest of the argument follows in due course.
3. The cause or explanation
of its [the universe’s] existence is something other than the contingent being
itself.
4. What causes or explains
the existence of this contingent being must be either other contingent beings
or include a noncontingent (necessary) being.
5. Contingent beings alone
cannot cause or explain the existence of a contingent being.
6. Therefore, what causes or
explains the existence of this contingent being must include a noncontingent
(necessary) being.
7. Therefore, a necessary
being (a being which, if it exists, cannot not exist) exists.
Premise 3 rests on the impossibility of something being a
causa sui. This must be distinguished from an individual being a
self-explainer in the sense that its existence is entailed by its essence since
there is a successful ontological argument for the existence of this
individual, even if we aren’t smart enough to give it. A self-explaining being
satisfies PSR but not PC, since the explanation of its existence is not a
causal one. Premise 4 follows from the application of the law of excluded
middle to premise 3. The truth of premise 5 becomes manifest once it is
realized that the contingent being in question is the universe, and thereby
includes every contingent existent. Were one of these contingent existents to
causally explain the existence of the universe, it would have to causally
explain, among other things, its own existence; but this is not possible, since
no individual can be a causa sui. Premise 5 is an obvious consequence of
4, for the reason just given. Since the universe, being itself a contingent
being, must have a cause, and this cause cannot be a contingent being, it must
be, or include, a necessary being. This is because every being is either
contingent, necessary, or impossible; and, obviously, an impossible being
cannot causally explain anything. The conclusion, 7, is an obvious logical
consequence of 6.
Why Reichenbach’s Argument Does Not Work
Three objections will be advanced against Reichenbach’s
argument: (1) the nontheist opponent of the argument is within her rights to
reject its PSR or PC, especially since no positive support is given for PSR or
PC by the cosmological arguer; (2) the argument violates the spirit and intent
of PSR in countenancing a God who has a brute, unexplained existence; and (3)
there is an unclosed gap between the necessary being that the argument
allegedly proves to exist and the God of traditional Western theism.
(1)
This objection applies to all of the traditional cosmological arguments
surveyed by Reichenbach. What they have in common is that each employs a
“strong version” of PSR because it is required for every fact of a certain type
that there actually is an explanation of it.1 A “weak version,” in
contrast, requires only that for every fact of a certain type it is possible
that there is an explanation of it. Within each of these two versions of PSR
distinctions can be made between weaker and stronger versions thereof. The strongest
version of the strong version of PSR requires that for every fact or true
proposition, without exception, there actually is an explanation of it.
Reichenbach’s version of the strong version of PSR is considerably weaker than
this, for it requires only that for every proposition that reports the
existence of a contingent being or the coming into existence of some being
there actually is an explanation of it.2 God, as conceived of by
Reichenbach, falls outside the purview of this version of PSR, since he is
neither a contingent being nor comes into existence. Rather, he is a necessary
being, not in the absolute, unqualified sense of necessary, but in an existentially
relative sense, namely, that if he exists, then he necessarily exists in the
sense that it is impossible that he either come into or go out of existence.
Why should the nontheist opponent of Reichenbach’s
argument grant his version of PSR? Although it is not the strongest version of
PSR, it still is quite strong, occupying a very exalted level in one’s wish
book, almost as high as that God exists. Reichenbach offers no direct
argumentative support for his version of PSR, but instead implicitly shifts the
onus on to his opponent to establish its falsity and, furthermore, attempts to
shoot down only one effort to do so, Hume’s.3
With what right does Reichenbach shift the onus? After
all, it is he who is advancing an argument for the existence of God that
crucially depends upon the acceptance of PSR, a principle that has been highly
mooted throughout the history of philosophy. Thus, it would seem that the
burden is on him to give positive argumentative support for it. I have not seen
among the multitude of cosmological arguers, past and present, any positive
argument for PSR. An example of what such an argument for PSR might look like
recently was offered to me by Alexander Pruss, with a grin on his face. Nothing
can exist without a sufficient reason for its existence; for then it would have
no connection with existence, and thus not exist. All that this argument shows
is that if a being has no sufficient reason for its existence, it has no
rational connection with existence, not that it has no connection at all. It
still could be connected with existence
in the sense of being a part of existence or instantiating the property of
having existence.
The
best that the cosmological arguer can do in support of PSR is to show that it
is pragmatically rational for an inquirer to believe it, since by believing
that everything has an explanation the believer becomes a more ardent and
dedicated inquirer and thus is more apt to find explanations than if she did
not believe this. This pragmatic sense of rational concerns the benefits that
accrue to the believer of the PSR proposition, as contrasted with the cognitive
or epistemic sense of rational that concerns reasons directed toward supporting
the truth of the proposition believed. Since Reichenbach’s argument attempts to
establish the cognitive rationality of believing that God exists, it cannot
employ a premise that concerns only the pragmatic rationality of believing some
proposition, such as PSR; for this would commit the fallacy of equivocation,
since “rational” would be used in both the pragmatic and cognitive sense. In
essence, it would be arguing that it is cognitively rational to believe a
proposition p because it is pragmatically rational to believe some proposition
q, from which p follows or which is needed for the deduction of p.
Not
only does Reichenbach offer no positive argumentative support for PSR and PC,
what little indirect support he gives is unconvincing. His indirect support
consists only in an attempt to refute Hume’s argument against PC. Simply put,
Hume argued that we can conceive of an uncaused event; and, since whatever is
conceivable is possible in reality, PC is false. Reichenbach’s rebuttal holds
that Hume “confuses epistemic with ontological conditions.” (p. 3) To be sure,
there is a distinction between what is conceivable and what could exist, the
former concerning the epistemic and the latter the ontological order.
Nevertheless, Reichenbach’s rebuttal is far too facile, for it fails to face
the fact that our only access to the ontological order is through the epistemic
order. The only way that we humans can go about determining what has the
possibility of existing is by appeal to what we can conceive to be possible.
Such modal intuitions concerning what is possible are fallible; they are only
prima facie acceptable, since they are subject to defeat by subsequent
ratiocination. They are discussion beginners, not discussion enders. What
Reichenbach has failed to do is to give Hume any reason why he should not trust
his prima facie modal intuition. And until such reasons are produced Hume has a
right to trust it. In philosophy we must go with what we ultimately can make
intelligible to ourselves at the end of the day, after we have made our best
philosophical efforts.
(2)
This is a variant of Schopenhauer’s objection to the cosmological argument as
being like a taxicab that we hire and then dismiss when we have reached our
destination. We begin by demanding, on the basis of PSR, an explanation for a
certain fact but when we arrive at our desired destination, God, we dismiss PSR
because we do not require an explanation for the fact that God exists. It is
just this sort of argument that invites the response from the precocious child,
“Yeah! And who created God!”
Reichenbach
has a ready response to this taxicab objection that is based on an alleged crucial disanalogy between his version of
the cosmological argument and those versions that are a suitable target for
this objection. Whereas both arguments begin with a demand to explain the
existence of some contingent being, his argument, unlike the objectionable
versions, terminates with a noncontingent explainer. That its existence is a
brute, unexplained fact is okay, since it is not possible that there be an
explanation for its existence, this being due to it being neither a contingent
being nor one that could begin or cease to exist. Pace Reichenbach, it will be
argued that it is possible for there to be an explanation for the existence of
his God and that even if this weren’t the case there still would be a violation
of the spirit and intent of PSR.
In
order to understand why, pace Reichenbach, it is possible for there to be an
explanation for the existence of his God, thereby making the fact that his God
exists fair game for any reasonable version of PSR, his notion of God’s
necessary existence must be made perspicuous. The sort of necessity that God
has, we are told, must not be confused with logical or absolute necessity,
terms that Reichenbach uses interchangeably, though the latter is preferable,
since not all absolutely or metaphysically necessary truths can be proven by
logic alone, such as that no object is larger than itself or red and green all
over. At first glance, the difference between the two senses of “necessary” is
that what is absolutely necessary is not relative to any world, whereas
Reichenbach’s sense of necessary is. If it is true in some possible world that
a being X necesssarily exists, then it is true in every possible world that X
necessarily exists. In contrast what is necessary for Reichenbach is
relativized to the actual world, it being possible for some being, in
particular his God, necessarily to exist in the actual world but not in every
other possible world. This seems to be the implication of his claim in step 7
that a necessary being is “a being which, if it exists, cannot not exist.” (p.
9) His definition of a “contingent being” as “a being which, if it exists, can
not-exist” also seems to relativize modalities to the actual world.(p. 9)
Whereas absolute modalities are world invariant, Reichenbach’s are relativized
to the actual world. As a consequence, his sense of necessity, unlike the
absolute one, is not subject to system S5’s basic axiom that if it is possible
that it is necessary that p, then it is necessary that p.
Although there is this
important difference between absolute and Reichenbachian necessity, it does not
adequately explain what the latter is. I believe that a perspicuous account of
Reichenbachian modalities will show that, appearances to the contrary, they are
analyzable in terms of absolute modalities. Let us look first at his definition
of a “contingent being” as “one which if it exists, can not-exist.” The first
thing to be noted is that this definition has the absurd consequence that every
absolutely impossible being, such as the object that is larger than itself, is
a contingent being, for it is true that if the building that is larger than
itself exists, then it can not-exist.” No doubt, Reichenbach would repair his
definition by restricting it to absolutely possible beings. The question, then,
is what his definition means by “can” when it says that an absolutely possible
being “can not-exist.” It looks like the “can” of absolute possibility. Thus,
his definition of a “contingent being” really says nothing more than that it is
absolutely possible for such a being to exist as well as absolutely possible
for it to not-exist. Reichenbach’s God, it might be added is a contingent being
in this absolute sense of contingent, since it is absolutely possible that he
exists and also absolutely possible that he does not exist.
It can be shown that
his existentially-relativized notion of a necessity also is analyzable in terms
of absolute modalities. What is meant by "cannot not exist” in his
definition of a “necessary being” as one which, “if it exists, cannot not
exist”? He explains this notion in terms of not having the possibility of
beginning or ceasing to exist, in which by “possibility” he would seem to mean
“absolute possibility.” Thus, the claim that God is a necessary being means
that it is absolutely impossible that God begin or cease to exist. Again, it
turns out that one of Reichenbach’s existentially-relativized modalities is
reducible to absolute ones.
Having
clarified what Reichenbach means by God’s having necessary existence, it can be
asked whether it is absolutely impossible that there be an explanation for the
fact that Reichenbach’s God exists. There is an explanation for God’s not
beginning or ceasing to exist based on it being a conceptual or metaphysical
truth that God’s nature precludes his doing so. But our question concerns
whether there could be an explanation for the fact that he exists at all.
Reichenbach so defines his God that the explanation cannot be in terms of God’s
nature, since he denies that his God necessarily exists in the absolute sense,
thereby precluding the possibility of explaining his existence via an
ontological argument. Although Reichenbach’s God lacks such necessary
existence, is it possible for some God-like being to have such necessary
existence?4 Reichenbach gives no argument against this possibility.
If it is possible for there to be a such necessarily existent God—one that
exists in every possible world—then it is possible that the existence of
Reichenbach’s God would be explained in terms of the causal efficacy of this
necessarily existent God based on what it wills. Thus, until Reichenbach
produces a telling argument against the possibility of there being a
necessarily existent God, he has no right to claim that it is not possible that
there be an explanation for the fact that his God exists. And even if he could
produce such an argument, it still would not follow that no explanation is
possible for the existence of his God, since it is possible that there is an
explanation for the fact that his God exists in terms of the causal efficacy of
some other equally contingent being. If Reichenbach’s God is omnitemporally
eternal, enduring throughout a beginningless and endless time, this second
being could also be omnitemporally eternal and sustain Reichenbach’s God
throughout this infinite time. Thus, it could causally explain his existence
without causing him to come into existence, that is, begin to exist in time,
something that is conceptually precluded in virtue of his very nature. Because
the causally explaining being does not have necessary existence, its existence
is yet to be explained. But this shows only that the proffered explanation is
not a final one, not that it is not a complete explanation. In conclusion,
Reichenbach’s claim that it is impossible for there to be an explanation for
the fact that his God exists shows a lack of imagination.
Since
his God is an absolutely contingent being, there should be at least the
possibility of explaining his existence, just as there is for any other
absolutely contingent being. That his version of PSR has the consequence that
every contingent being other than his equally contingent God have an
explanation makes his version of the cosmological argument a fit target for the
taxicab objection. His version of PSR violates the underlying spirit and intent
of the strong version of this principle, which, at a minimum, wants every
absolutely contingent being to have an actual explanation for the fact that it
exists. That Reichenbach’s God has the wondrous property of lacking the
possibility of beginning or cease to exist is not itself sufficient grounds for
making it an exception to PSR's demand that every absolutely contingent being
have an actual explanation.
(3)
The conclusion of Reichenbach’s argument, it will be recalled, is:
7. Therefore, a necessary
being (a being which, if it exists, cannot not exist) exists.
Reichenbach rightfully claims
that the necessary being in 7 does not provide a naturalistic explanation for
the existence of the universe, but then dubiously adds that “What remains is a
personal explanation in terms of the intentional acts of some supernatural
being with properties such as eternality and aseity.” (p. 11) This imputation
of eternality, understood in the timeless sense, and aseity to this being
appears to be a piece of prestidigitation of a similar ilk to St. Thomas’s
terminating each of his Five Ways with the claim “et hic dicimus Deum”--and
this we call God. This papers over the gap problem of how we are to establish
that his unmoved mover, first cause, etc. has all of the other required divine
attributes, among which are omnipotence, benevolence, and omniscience. This
problem infects every traditional cosmological argument. Reichenbach does
nothing toward closing the gap between the necessary being in 7 and the God of
traditional Western theism. Thus, he has no justification for ending his paper
with the claim that his argument, along with other cosmological arguments,
“provide a justification for theistic belief,” if what is in question is
traditional Western theistic belief. (p. 24) In fairness to the cosmological
argument, it must be pointed out that even with an unresolved gap problem, it
accomplishes something of great significance, if successful, namely
establishing the existence of a quite wondrous supernatural being who is the
causal explainer of the existence of the universe. To close the gap arguments
are needed to show that the being who is the causal explainer of the existence
of the universe has all of the other required divine perfections. Herein it
will be necessary to make use of a variety of different inductive arguments
based on the overall goodness of the universe, such as those from natural
purpose and widespread law-like order and simplicity.
II. A Sketch of a Cosmological Argument that Works
The following is a new argument for the existence of a
being who, if not the super-deluxe God of traditional Western theism, is at
least a close cousin in that it too is capable of playing the role in the lives
of working theists of a being that is a suitable object of worship, adoration,
love, respect, and obedience. Unlike the super-deluxe God, the God whose
necessary existence is established by my argument need not essentially
have the divine perfections of omnipotence, omniscience, and
omnibenevolence. Furthermore, he need not even be contingently omnipotent and
omniscient, just powerful and intelligent enough to be the supernatural
designer-creator of the very complex and wondrous cosmos that in fact confronts
us. Hopefully, his benevolence can be taken to be unlimited. My reasons for
preferring to work with this more limited God is not just that I am able to
prove his existence but not that of the super-deluxe one. It also involves, as
will emerge later, the ability of the concept of a finite God to get around
certain difficulties that confront the traditional conception of God as an
absolutely perfect being.
The
new argument that I will now sketch is not just a cosmological argument, but a
cosmological cum ontological cum teleological argument. The main argument is a
watered down version of the S5 modal version of the ontological argument,
followed by a cosmological argument for its only controversial premise, and
concluding with a collection of teleological arguments that attempt to close
the gap between the absolutely necessary being whose existence is proven by the
former arguments and the God of traditional Western theism.
The Main Argument
1. If it is possible that it
is necessary that there exists a supernatural being who is a very powerful and
intelligent designer-creator of the universe, then it is necessary that there
exists a supernatural being who is a very powerful and intelligent designer-creator
of the universe.
2. It is possible that it is
necessary that there exists a supernatural being who is a very powerful and
intelligent designer-creator of the universe.
3. It is necessary that there
exists a supernatural being who is a very powerful and intelligent
designer-creator of the universe. By
modus ponens from 1 and 2
Premise 1 is a substitution
instance of the axiom of S5 that if it is possible that it is necessary that p,
then it is necessary that p. This is a special case of the general principle
that a proposition’s modal status—its being necessary, possible, or impossible—is
world invariant. A being has necessary existence if and only if it is necessary
that it exists. Such a being exists in every possible world.
It might be wondered why I did not avail myself of the
robust version of the S5 modal ontological argument, in which the first premise
asserts that it is possible that it is necessary that there exists a being that
is essentially omnipotent, omniscient, omnibenevolent, and so on for all the
other essential divine perfections. Why did I not cast my Main Argument in
terms of this absolutely perfect being rather than in terms of a finite God who
does not have any of these omni-properties, possibly with the exception of
omnibenevolence, no less have them essentially? It is not just that I am able
prove the existence of my less than absolutely perfect deity but cannot do so
for the superdeluxe model but rather that the latter is an impossible being.
The reason is that this being, since necessarily existent, exists in every
possible world and is at its greatest greatness in every one of them, since it
essentially has all of its omni-perfections. But this has the absurd
consequence that certain things that are possible are not possible. For
example, it is possible that there exists a completely gratuitous, unjustified
evil, that is, an evil that is incompatible with the existence of God. But,
since the superdeluxe God exists in every possible world, in no possible world
will there be such an unjustified evil, and thus it is not possible that there
exists a completely gratuitous, unjustified evil.5 The danger of making
God too perfect is that it makes him an impossible being and thus not perfect
after all. But the danger of making God finite, as my argument does, is that it
might make him too finite and thereby not a suitable object of worship,
reverence, love, and obedience
Since
my main argument is a deductively valid modus ponens argument, the only way
that it can fail is by having a false premise. Assuming that one is willing to
grant the S5 axiom that underlies premise 1, the only possibly dubious premise
is 2, and dubious it is indeed. For, whereas the biblical fool was happy to
grant the possibility premise in St. Anselm’s ontological argument—that it is
possible that there exists a being who essentially has all of the divine
perfections—he would have to be not just a fool but a complete schmuck to grant
the possibility premise, 2, of the Main Argument. For he should not give a
consent that is not an informed consent, and an informed consent to 2 requires
knowing what it is meant by its nested operators—it is possible that it is
necessary that—namely, that it is subject to S5’s axiom. But once the fool
realizes this he will withhold his consent, since he will rightly see premise 2
as begging the question against the nontheist opponent of the argument.
Plainly, an argument is needed for 2, and it is the purpose of my Subsidiary
Argument, which is a new version of the cosmological argument, to do just that.
The Subsidiary Argument
My argument will make use of some technical terms that
need to be clarified at the outset. A possible world is a maximal, compossible
set of propositions, maximal because for every proposition either it or its
negation is a member of the set and compossible because all of the propositions
in the set could be true together. A contingent proposition or being is both
possibly true or existent and possibly false or nonexistent. Since my concern
is with establishing that the proposition that God exists is a member of the
actual world—that maximal, compossible set of propositions all of whom are
actually true, I will confine my attention to the actual world. There is a
proper subset of the actual world consisting in all the contingent propositions
in the actual world that do not report the action of a necessary being.
Corresponding to this proper subset is the Big Conjunctive Fact that has as its
conjuncts the members in this proper subset. In order to avoid the absurdity
that the Big Conjunctive Fact is one of its own conjuncts, given that the Big
Conjunctive Fact is itself a contingent proposition that does not report the
action of a necessary being, it must be forbidden for the Big Conjunctive Fact
to contain a conjunct that is truth-functionally equivalent to a simpler one,
as for example the conjuncts p and (p and p), or p and (p or p). The universe
or cosmos is comprised of all the contingent beings and the events in which
they participate that are referred to and described by the propositions in the
Big Conjunctive Fact. It is the universe that renders these propositions true,
serving as their real-life verifiers.
The general strategy of my of my argument is to adopt the
following weak version of PSR:
PSRw. For every true
proposition, p, it is possible that there is an explanation for p.
and then show that the only
possible explanation of the actual world’s Big Conjunctive Fact is in terms of
the creative actions of a necessary supernatural being who is possessed of very
great power and intelligence. That it is possible that there is an explanation
of the Big Conjunctive Fact in terms of the causal efficacy of such a necessary
being entails that it is possible that it is necessary that there exists such a
being, which is a stylistic variant of the proposition that is to be proven:
2. It is possible that it is
necessary that there exists a supernatural being who is a very powerful and
intelligent designer-creator of the universe.
By being able to make do with
a weak version of PSR my new cosmological argument has a distinct advantage
over past cosmological arguments in that it requires only the possibility, not
the actuality of an explanation for every fact.6 Thus, it presents a
less vulnerable target to its opponent, exacting a greater price for rejecting
its version of PSR. The principle of Minimal Ordinance enjoins us to make use
of the weakest premises possible for establishing some conclusion, since thereby
we render our premises less vulnerable to challenge.
There is, however, an ambiguity in PSRw between a weak
and a strong version that must be faced and resolved at the outset. The ambiguity
concerns whether a proposition that is a possible explainer of a proposition
must be a member of the same world as it is. The weak version of PSRw holds
that
PSRww. For any possible world
w and any proposition p, if p is a member of w, then it is
possible that there is a world w1 and a proposition q such
that q is a member of w1 and q explains p
in w.
In which w need not be
identical with w1. The strong version holds that
PSRws. For any possible world
w and any proposition p, if p is a member of w,
then it is possible that there is a proposition q such that q is
a member of w and q explains p in w.
It will turn out that my
argument requires the adoption of PSRws. But this stronger version of the weak
version of PSR still is considerably weaker than any strong version of PSR,
thereby giving my cosmological argument an important advantage over all those
cosmological arguments that require a strong version of PSR.
With these preliminaries out of the way, my argument can
now be given. The actual world, which will be called a, is a set of
propositions, namely all the propositions that would be true were a to be
actualized, which we are assuming actually to be the case. Let us give the name
“p” to the Big Conjunctive Fact that has as a conjunct every contingent
proposition in a that does not report the action of a necessary being. Thus, it
is true ex hypothesis that
4. p is the Big
Conjunctive Fact for a.
From 4 it follows, in
accordance with PSRws, that
5. It is possible that there
is a proposition q such that q is a member of a and q
explains p in a.
Surprisingly, from 5 it can
be deduced that
6. There is a proposition q
such that it is possible that q is a member of a and q
explains p in a.
I say surprisingly because,
in general, it is fallacious to reverse the order of 5’s modal operators, that
is, to deduce that there is an X such that it is possible that X is F from the
proposition that it is possible that there exists an X who is F. For example,
it could be possible that there is someone who flies to Venus without it being
true that there exists someone who possibly flies to Venus, since the former
proposition, unlike the latter, could be true even if there never exists such
an individual. The reason why it is not fallacious to deduce 6 from 5 is
because propositions, like other abstract entities such as properties and
numbers, enjoy necessary existence, existing in every possible world. Thus, if
it is possible that proposition p exists (i.e., the proposition that p
exists is a member of some possible world), then proposition p exists
(i.e., the proposition that p exists is a member of the actual world).
It might be objected that the possible proposition, q,
that is the explainer of p in a need not be a member of a, as is
allowed by PSRww; for, since q is possible, it is possible that q
is a member of a, even if it isn’t. This objection confounds
(a) The proposition that it
is possible that q is a member of a.
with
(b) It is possible that
proposition q is a member of a.
Proposition (a) is true in
virtue of the fact that a proposition’s modal status is world invariant; thus,
if p is possible in one world it is possible in every world. Proposition
(b), however, is not entailed by (a), since possibly p could be a member
of a given world without it being possible that p is a member of that
world. This is due to the fact that a possible world is a set of propositions
and that it is not possible that a set have different members than it in fact
has. Its very identity is determined by its membership: Were it to have
different members, it would be a different set. Therefore, unless q
actually is a member of a, it is not possible that q is a member of a,
even though the proposition that it is possible that q is a member of a.
Furthermore, if it is possible that q is a member of a, then q
is a member of a. Thus, it follows from
7. There is a proposition q
such that it is possible that q is a member of a and q
explains p in a.
that
8. There is a proposition q
such that q is a member of a and it is possible that q
explains p in a.
We are getting close to
deriving premise 2 of the Main Argument, but, before this can be done, more
must be said about what sort of proposition q is.
Recall
that every contingent proposition in a that does not report the action
of a necessary being is a conjunct in q. Since q is a member of a
and possibly explains p in a, q cannot be a contingent
proposition that does not report the action of a necessary being. Were q
to be such a proposition, it would be a conjunct in p and thus would
have to explain, among other things, itself. But such an explanation, as
Reichenbach has ably shown, would be viciously circular.
We need to know more about
what sort of a proposition q is than just that it is not a contingent
proposition that does not report the action of a necessary being. The open
possibilities are that q is either a necessary proposition or a
contingent proposition that reports the action of a necessary being. The former
does not seem up to the task of explaining p, since the mere existence
of a necessary being, even if it were God, could not explain why all of the
contingent propositions in p are true together. In other words, it would
not explain why there exists the universe that is reported by the members of p.
It must be something that a necessary being does that explains p This
necessary being cannot be an entity without intelligence, power, and will, such
as a number or a Platonic form. The only type of explanation that we can
imagine or conceive of for p, given that p cannot have an explanation whose explanans contains at
least one contingent proposition that does not report the action of a necessary
being, is a personal one in terms of the intentional actions of a necessary
being. Of course, it is possible that there are forms of explanation that we
are not capable of imagining or conceiving of, but, as already has been argued,
in philosophy we must go with what we can conceive of after we have made our
best effort. Given that, relative to our powers of conception, the explanation
for p must be a personal one in terms of the intentional action of a
necessary being, it is reasonable to assume that this action is a contingent
one, and thus q is itself a contingent proposition.
The
conclusion that we must draw is that q is the proposition that there
exists a necessary supernatural being who is the causal explainer of the
universe, that is, p, the Big Conjunctive Fact. Given the wondrous
complexity and fine tuning of the universe, this being must be a very powerful
and intelligent designer-creator. Thus, q is identical with the
proposition that that there exists a necessary supernatural being who is a very
powerful and intelligent creator-designer of the universe, and therefore the
latter can be substituted for the former.
From step
8. There is a proposition q
such that q is a member of a and it is possible that q
explains p in a.
it can be deduced that
9. It is possible that there
exists a necessary supernatural being who is a very powerful and intelligent
designer-creator of the universe.
This is because of the
following three facts: any proposition that is a member of the actual world, a,
is true; we are licensed to substitute for q the proposition that there
exists a necessary supernatural being who is a very powerful and intelligent
creator-designer of the universe, since they are one and the same proposition;
and the universe is what is reported by p and thus whatever explains one
explains the other. Given that if a being is necessary, then it is necessary
that it exists, it can be deduced from 9, in accordance with S5’s axiom, that
2. It is possible that it is
necessary that there exists a supernatural being who is a very powerful and
intelligent designer-creator of the universe.
QED
I apologize for the looseness and sloppiness of this
deduction of 2. For the sake of brevity, many crucial steps have had to been
omitted. For a rigorous logistical deduction of 2 (containing 32 steps!) I must
refer the reader to my essay, “A New Argument for the Existence of God: One
that Works, Well, Sort Of.”7
Objections
I
will now consider some objections with the hope that this will help to deepen
the reader’s understanding of the significance of my argument, making clear
just what it does and does not accomplish.
The
Explanation Is Agglomerative Objection.
A crucial step in my argument was the claim that The Big Conjunctive Fact in a
given world—the conjunction of all the contingent propositions in that world
that do not report the action of a necessary being—is explainable only by a
proposition that is not a member of the conjunction. It could be objected in
the name of Hume that if the conjunction were infinite, with each conjunct
being explained by another conjunct, the entire conjunction would thereby be
internally explained. This assumes that explanation is agglomerative, meaning
that it is closed under conjunctive introduction: If there is an explanation
for p and another explanation for q, then there is an explanation for the
conjunction (p and q). Pace the principle of the agglomerativity of explanation,
it is possible that it is a mere coincidence that p and q are true together,
even when each of them has some explanation. It also is possible that there is
a common cause that explains their conjunction.8
The
Taxicab Objection. A major virtue of
my argument is that it, unlike past cosmological arguments, including
Reichenbach’s, escapes this objection, because my explanation for The Big
Conjunctive Fact is in terms of a proposition that ends the regress of
explanations. The proposition that there is a very powerful and intelligent
necessary being that causes the existence of the cosmos in world a (or brings
it about that The Big Conjunctive Fact for a is true) is a self-explaining
proposition in spite of it being a contingent proposition, provided it is added
that it freely does so. The reason for this is that a necessary being is one
whose existence can be explained by an ontological argument, even if we cannot
give it, and that a being freely performs an action, such as causing world a’s
cosmos to exist, stands in need of no further explanation, at least on the
Libertarian Theory. Once it is said that the being freely does it that explains
his action. Thus, the proposition that some necessary being freely does action
A is a regress-of-explanation ender.
But,
it might be objected: Why assume that my necessary being freely causes the
cosmos to exist? The reason is that it is hard to understand how a very
powerful and intelligent supernatural being who is a cosmos-causer would not
act freely. What could possibly coerce or compel him to act as he does, for he
determines and controls every feature of the cosmos?
The Unintelligibility of Theistic Explanations.
There are philosophers who find unintelligible the notion of a purely spiritual
being freely causing there to exist a cosmos by his will because there is not
the required relation of statistical relevance between his free effort of will
and its effect, the resultant cosmos. I cannot in this paper do justice to this
objection, since a proper response to it would have to defend the coherence of
theism against this and many similar types of objections, such as that the
theistic explanation for the existence of the cosmos does not enable
predictions to be made and thus is no explanation.
The general strategy for a response to the
incoherency-of-theism-objection is to charge it with employing a
question-begging scientistic premise, which I will call “The Legislativeness of
Scientific Contexts” principle. This principle holds that the features that
inform the use of a concept in a scientific context are legislative for the use
of this concept in every context, any use that does not incorporate them being
unintelligible. Thus, Grunbaum finds through his analysis of the use of the
concept of causation in scientific contexts that it involves a relation of
statistical relevance between the cause and its effect, and thereby demands on
the basis of the principle of the Legislativeness of Scientific Contexts that
every use of the concept of causation have this feature. Since theistic uses of
the concept of causation do not, he charges them with being unintelligible. One
has only to state this principle in order to defuse the
unintelligibility-of-theistic-explanations objection that is based on it. For
the principle is not one that is vouchsafed by science. Rather it is a
metaphysical thesis that fails to find adequate argumentative support and can
rightly be charged by the theist with begging the question.
The Nonpersonal God Objection. Phil Quinn, in
correspondence, has questioned my claim that the only type of explanation that
we can imagine or conceive of for the Big Conjunctive Fact, p, in the actual world, given that it cannot
have an explanation whose explanans contains at least one contingent
proposition that does not report the action of a necessary being, is a personal
one in terms of the intentional actions of a necessary being. He writes:
I agree
that the necessary being cannot be a number or Platonic form. Nor, I would add,
can it be the Plotinian One, from which the cosmos emanates of necessity. I
also agree that it cannot be without power. But I think it can be without
intelligence or will. I can conceive of explaining r in the following
way: There is an impersonal necessary being, rather like the Brahman of advaita
Hinduism, that generates the cosmos by means of blind but indeterministic
mechanical causation. So I am inclined to think that even if your Subsidiary
Argument, as presented in the body of the paper, goes through, it will not
yield the minor premise of your Main Argument.
There are several ways of attempting to meet this
interesting objection. First, I could concede the objection and work with a
more generic brand Deity who is a common denominator of the different
cosmos-explaining necessary supernatural beings. My argument, then, would prove
the existence of a necessary supernatural being of considerable power who is
the cause, though not necessarily in a personal manner, of the cosmos. This is
no mean feat; however, I don’t think I have to concede to Quinn’s objection. In
the first place, the Brahman of the advaita is not a necessary being in the
sense that is relevant to my argument, namely a being the concept of which
explains its existence. Furthermore, it is dubious that the purported explanation
of the cosmos in terms of the blind, indeterministic activity of this
impersonal force is any better explanation of the existence of the cosmos than
that in terms of a mystical One out of which the actual cosmos emanates. This
cosmos displays considerable lawlike regularity and simplicity, as well as
remarkable fine-tuning of its physical constants, all of which goes unexplained
by an impersonal “explanation.”
The
My-Argument-Doesn’t-Do-Enough Objection.
It was the aim of my argument to escape the closing of the gap problem that has
infected past cosmological arguments, the unwarranted move from a conclusion
that there exists a first mover (cause, etc.) to the claim that this being is
God, that is, has all of the divine perfections. But in avoiding the Scylla of
the gap problem I may have wrecked on the Charybdis of proving the existence of
a being who falls far too short of the divine mark.
One aspect of the problem concerns whether my “God” is
powerful and intelligent enough to be a suitable object of worship and
adoration. Given the incredible complexity and wonderfulness of the actual
cosmos, I am not too worried about this problem, since any being who is capable
of designing and causing this cosmos is sufficiently awesome in his power and
intelligence to be a suitable object of worship and adoration by the working
theist. That this “God” need not be either omnipotent nor omniscient, no less
essentially so, will worry the great medieval theists, who were after bigger
game, but it should not render him unserviceable for the needs of ordinary
believers. Furthermore, by having a finite God, it might make it a lot easier
to construct plausible theodicies, such as were available to Plato in the
Timaeus, but this is a direction that cannot be pursued here.
The most serious problem, however, concerns the moral
attributes of my “very powerful and intelligent necessary being that causes the
existence of the cosmos in the actual world.” This issue does not concern the
existence of this being but rather whether it is a suitable object of worship,
adoration, and obedience. If I cannot show that this being is at least a very
good being, my argument may very well have created a Frankenstein.
To
begin with, my creator God is not claimed to be essentially omnibenevolent,
which, I take it, is a virtue of my argument. In the first place, it saves
God’s freedom, which was required to meet the taxicab objection. Most
important, it results in God not being omnibenevolent in every world in which
he exists. This is important for the reason given earlier, namely that since he
exists in every world it would not be possible for there to be a morally
unjustified evil in any world, assuming, of course that he also is essentially
omnipotent and omniscient. But plainly it is possible for there to be a purely
gratuitious evil. What matters to the working theist is not whether it is
logically possible that God do what is morally wrong but whether he is capable
of doing so in the actual world, in which capable is understood in terms of
what a being has the capacity, knowledge, and opportunity to do. God could be
said to be incapable in the actual world of doing wrong in the sense that he
could not get himself to do so, that he is above temptation, that we can place
absolute confidence in him. What he does in other possible worlds is
unimportant to the working theist.
But
this still leaves it open whether we have good reason to think that my “very
powerful and intelligent necessary being that causes the existence of the
cosmos in the actual world, a,” is benevolent in the actual world. It is here
that my argument becomes quite vulnerable. To meet this problem I’ll have to
marshal all of the extant theodicies for God’s permitting all of the known
evils of the world. My task is made easier because my God might be finite and
thus possibly could use the excuse of Plato’s Demiurge. But this raises in turn
the falsifiability problem. How finite is my God? I’ll have to leave it to
those more skilled in apologetics to take up the cause here. The best that I
think can be done is to argue that the actual cosmos is overall a good one in
that it is better that it exist than that it not exist. And, if there is an
infinite regress of worlds in respect to goodness, as seems reasonable, God
cannot be faulted for not actualizing the best possible world. But after all is
said and done, the unqualified goodness of my God in the actual world remains
unproven and must ultimately be a matter of faith.
In
conclusion, I believe that my argument goes quite some way to justifying
theistic belief, maybe even making it more likely than not that my finite God
exists. Although my argument justifies theistic belief, it does not make it
rational for someone to believe in theism on its basis alone. The reason for
this is that an essential requirement for it being rational to have a theistic
belief is that the believer has had some experiential awareness of God or of
God’s presence in the cosmos. She need not have had a direct, nonsensory
perception of God, no less a mystical experience of at least partial union with
God, but at least she must have had experiences in which she perceived worldly
items as being God caused. William James, and following him William Alston,
have been quite right to stress the central role that religious experience
plays in religious belief. By driving a wedge between justification and
rationality, it is shown how it is possible for someone to accept my argument
and yet be an atheist. Recall in this connection Bertrand Russell’s claim that
at one time he accepted the ontological argument but still persisted in his
agnosticism. If I am right, there is nothing absurd about this.9
Richard M. Gale
University of Pittsburgh
1. By “fact” is meant a true
proposition, a proposition being an abstract entity that can serve as the
bearer of truth-values and modalities,
as well as the object of an intentional attitude, such as believing,, and the
meaning or sense of a sentence. For reasons that will not be gone into here, it
is such abstract propositions that get filled in for the blank spaces in
“___explains (or causes)___.”
2. “What is contingent or
what comes into being requires a sufficient reason for why it exists or comes
into being.” (p. 2)
3. Thus I was surprised to
see Reichenbach subsequently write, on p. 18, “We have already discussed and
defended 8, which is a version of the Causal Principle that underlies all
cosmological arguments.” This is quite misleading, since previously he gave
only a weak indirect defense of this principle.
4. And, in what follows, I
will argue that this is not only possible but actually the case.
5. This is argued for at
length in Chapter 5 of my book On the Nature and Existence of God
(Cambridge: Cambridge University Press, 1991).
6. The source of inspiration
for my employment of the weak version of PSR is Duns Scotus, who used it in an
ontological argument that went as follows. For every fact, it is possible that
there is an explanation for it. Nothing can prevent God’s existence. Therefore,
it is not possible that there is an explanation for the fact that God does not
exist. If God were not to exist, it would follow that it both is and is not
possible that there is an explanation for this fact.
7. Forthcoming in The
Rationality of Theistic Belief, edited by Godehard Bruntrump, on Kluver
Academic Press.
8. This is argued for at
length in Chapter 6 of my On the Nature and Existence of God. Alexander
Pruss, in his excellent article, “ the Hume-Edwards Principle and the
Cosmological Argument,” forthcoming in the International Journal for the
Philosophy of Religion, gives additional arguments for why explanation is
not agglomerative.
9. In the discussion of my
argument at a conference on “The Rationality of Theistic Belief” at the Munich
School of Philosophy on May 27, 1998, Uwe Meixner raised a devastating
objection to my argument; but, fortunately, Alvin Plantinga was in the audience
and immediately showed me a way of avoiding it. I am deeply indebted to both of
these persons.