COSMOLOGICAL AND DESIGN ARGUMENTS

Alexander Pruss and Richard M. Gale

Unlike the ontological argument, which appeals only to highly sophisticated philosophers of a mathematical bent of mind, cosmological and design arguments figure prominently in the argumentative support that everyday working theists give for their faith. The reason for this broad pastoral appeal is that these arguments begin with commonplace facts about the world and then, by appeal to principles that look plausible, establish the existence of a being who, while not shown to have all of God’s essential properties, properties that God must have to exist, is at least a close cousin of the God of traditional Western theism. Our plan is to begin with a preliminary botanization of these arguments, indicating their similarities and differences, and then discuss each of them separately, giving prominence to the many different forms they take.

Preliminary Botanization

Each of the two arguments begins with a contingent existential fact. A contingent fact is a true proposition that has the possibility of being true and the possibility of being false, in which possibility is understood in the broadly logical or conceptual sense. By extension a contingent being is one who has both the possibility of existing and the possibility of not existing, with a necessary being not having the possibility of not existing. The arguments differ with respect to the type of existential fact that they select. For design arguments it will be a fact that reports some natural object or process that displays design, purpose, function, order, and the like. It might be the fact that there is life, self-replicating organisms, consciousness, conscience, law-like regularity and simplicity, natural beauty, and apparent religious miracles. In contrast, a cosmological argument’s existential fact does not have any of these sort of normative features. It might be the fact that there exists a total aggregate of contingent beings (the universe), or maybe that there exists at least one contingent being, or that one object depends upon another for its existence.

The two types of argument also differ in the way in which they go from their initial contingent existential fact to the existence of a supernatural God-like being who is the cause of this fact. A cosmological argument, typically, demands a cause of this fact in the name of the Principle of Sufficient Reason (hereafter PSR), which is suitably tailored so that every fact of this kind actually has an explanation. This is followed by an explanatory argument to show that the only possible explanation for this fact is in terms of the intentional actions of a God-like being. Thus, a cosmological argument standardly has the following three components.

1. a contingent, nonnormative existential fact

2. a version of PSR that requires that every fact of this kind has an explanation

3. an explanatory argument to show that the only possible explanation of this fact is in terms of the intentional actions of a supernatural, God-like being.

In contrast, the typical design argument does not demand an explanation for the initial contingent existential fact on the basis of some version of the PSR but instead employs principles of inductive reasoning to infer that it is highly probable that this fact is caused by a supernatural, God-like being. These principles might involve principles of analogical reasoning or abductive inference (inference to the best explanation). Thus, the typical design argument has the following three components.

1’. a contingent, normative existential fact

2’. some principle of inductive reasoning

3’. an explanatory argument to show that the probable explanation of this fact is in terms of the intentional actions of a supernatural, God-like being.[1]

It is important to stress that these components comprise only the typical design argument, for there are versions that do not employ 2 and 3. Some design arguments do not induce but instead deduce from the fact reporting some occurrence of natural design that there is a supernatural designer-creator of this occurrence, it supposedly being an analytic truth that something displaying design or purpose must have a designer or purposer. This does not make for an effective argument, since its opponent will be within her rights to charge its existential fact component with begging the question. There are Thomistic type design arguments that also attempt to deduce the theistic conclusion from the initial existential fact but do not appeal to this trivializing analytic truth but instead some high-level metaphysical principles.

Cosmological Arguments

With these preliminaries out of the way, we can begin our survey of the different types of cosmological arguments. Saint Thomas Aquinas presented Five Ways of proving the existence of God, the first three of which are versions of the cosmological argument. The First Way begins with the contingent fact that one object is moved by another, the Second that one thing depends for its existence upon the causal efficacy of a contemporaneous being, and the Third that there exists a contingent being. These are commonplace observational facts that only a complete sceptic about our senses would want to challenge. The explanatory arguments in the First and Second Ways are based on the impossibility of there being respectively an infinite regress of objects simultaneously being moved by other objects or objects depending for their existence upon the simultaneous causal efficacy of another being. These regresses, therefore, must terminate with a being who is capable respectively of moving another object without itself being moved by another or causing the existence of something without itself being caused to exist. Thomas then identifies this first mover or cause with God on the basis of our common ways of speaking about God -- “et hic dicimus Deum” -- thereby papering over a serious gap problem, since his arguments do not establish that these beings have all of the essential divine attributes.

The intuition underlying Thomas’s rejection of the possibility of an actual infinity of simultaneous movers or causers is far from obvious, especially since he did not think it impossible to have an actual past infinite regress of nonsimultaneous causes, as for example an actual infinite regress of past begetters. We will make an attempt to draw out his intuition in a way that gives some plausibility to it. The causal relation in a series of simultaneous causes or movers involves transitivity in that if X simultaneously moves (causes) Y and Y simultaneously moves (causes) Z, then X moves (causes) Z. Nonsimultaneous causation is not transitive, since, even though you were begot by your parents and they in turn were by begot by their parents, you were not begot by the latter.

One reason that might be given for the impossibility of an actual infinite regress of causes or movers is that if there were such a regress, there would be no member of the regress that could be held to be morally responsible, a fit subject of either praise or blame, for the initial event or object in the regress. But this can’t be the right reason, since not all causal explanations are forensic in the sense of giving an individual who is to be praised or blamed for the effect. Maybe Thomas’s underlying intuition can be fleshed out by considering these two examples. In one a group of boys attempt to get into the movies free by having each boy point to the boy behind him as he enters the theatre and when the ticket taker stops the last boy in the group for the tickets he claims not to know who these other boys are. (Richard Gale did it but Alexander Pruss did not, since he grew up in Communist Poland.) The last boy has to pay for himself but all the others get in free. Now suppose that the regress of boys pointing behind themselves to another boy is infinite. Plainly, the theatre owner would not be happy with this arrangement, since he would never get paid, just as you would never succeed in cashing a check if it were covered by a bank account that in turn was covered by another and so on ad infinitum. A system of credit, like a succession of boys entering a theatre, must terminate with some actual cash. A second example involves a train of cars that simultaneously push each other, such that the first car is simultaneous moved by a second, and the second by a third, and so on ad infinitum. If the regress of movers were infinite, there would be no explanation of where the oomph, the energy, the power to move, came from.

There is an implicit appeal to a version of the PSR to the effect that something cannot come out of nothing. This can be made clearer by considering a circle of causes. Thomas ruled this to be impossible for the same intuitive reason that he proscribed an infinite regress of simultaneous movers or causes. Imagine that you meet someone who looks like you would look in ten years. She claims to be your future self and to have traveled ten years backward in time in order to give you instructions on how to build a time machine. Subsequently, you build one and then travel ten years backward in time so as to inform your past self about how to build a time machine. The intuitive grounds for Thomas’ rejection of the possibility of this closed causal loop is that it violates the PSR, for there is no answer to the question of from whence came the knowledge of how to build a time machine. Similarly, there is no answer to the question of from whence came the power to move an object or causally sustain its existence in the case of an infinite regress of simultaneous movers or causers.

The Third Way begins with the unexceptionable contingent existential fact that there now exists at least one contingent being. Can some version of the PSR be employed so as to deduce that there exists a necessary being that causes the existence of this contingent being? A contingent being has the possibility of not being, and thus given an infinite number of times, either through an infinitely extended past or a past time interval that is comprised of an infinity of moments of time, this possibility will be realized at some past time. Each moment is like a roll of the dice, an opportunity for this possibility to be realized. The PSR tell us that something cannot come out of nothing, so there has to be a cause of this being’s coming into existence at this past time. Therefore, something had to cause this being to come into being out of nothing. But why couldn’t this cause be itself a contingent being and it, in turn, be caused to begin to exist by an even earlier contingent being, and so on ad infinitum? Thomas’ answer as to why this regress of contingent beings is impossible seems to commit an egregious quantificational blunder. For he says that if there were to exist only contingent beings, then, since for each of them there is a past time at which it doesn’t exist, there is a past time at which each one of them does not exist. And, if there ever were nothing, then, given the PSR, nothing would subsequently exist, which contradicts the patent existential fact that there now exists at least one contingent being. This argument seems to commit the same howler as is committed by inferring from the fact that for every woman there is a man that there is a man who is for every woman (talk about polygamy!). In logistical terms it is (x)($y)xRyÉ($y)(x)xRy. But it is hard to believe that a great philosopher committed so obvious a blunder. With a little charity and imagination something interesting can be made out of the Third Way, but we shall not attempt to do so here.

But it is hard to believe that a great philosopher committed so obvious a blunder. With a little charity and imagination something interesting can be made out of the Third Way. If the PSR is not to be violated, there must be an infinite number of contingent beings, assuming that there is no necessary being. For if there were at present a finite number of contingent beings, there would be a past time at which there is nothing, given that for each of these there is a past time at which it does not exist. Herein there is no quantificational fallacy. There must, then, be an infinite number of contingent beings, such that each of these beings is brought into existence at some past time by another contingent being.

At this point the argument can be developed in two different ways. The first is to argue that there cannot simultaneously exist an infinite number of beings. Appeal could be made to various Hilbert’s Hotel paradoxes that break out from this being the case. These paradoxes will be spelled out when we next consider the Kalan cosmological argument. The second way to proceed is to demand in the name of the PSR an explanation for the infinite aggregate of dependent beings. The causal explainer cannot be a member of this aggregate; for then it would have to explain its own existence, which is impossible. But since all existent beings are assumed to be contingent, there are no beings outside the aggregate and thus the existence of the infinite aggregate goes unexplained, thus violating the PSR. On this interpretation the Third Way turns out to be identical with Clarke’s cosmological argument, which will be presented right after the Kalan cosmological argument.

The Kalan Cosmological Argument of the medieval Islamic philosophers, which has been defended in recent times by William Lane Craig, also invokes the impossibility of infinite regress but in a different way that Thomas did in his first Two Ways. It selects as its contingent existential fact that there now exists a universe – an aggregate comprised of all contingent beings. It then argues that the universe must have begun to exist, for otherwise there would be an actual infinite series of past events or time, which is conceptually absurd.[2] Since something cannot come out of nothing, there had to be a cause for the universe coming into being at some time a finite number of years ago. And this cause is identified with God, which again occasions the gap problem. Notice that the version of the PSR that is appealed is a restricted and thus less vulnerable version of the PSR; for whereas the unrestricted version requires explanation for every thing that exists or fails to exist, the former requires an explanation only for a being’s coming into existence.

Just why is it impossible for there to be an actual infinity of past events or times? The answer is not obvious. Thomas, for one, did not think it to be impossible. Alex do your stuff here. Might relate to the Third Way. See above promissory note.

Probably the most powerful of the traditional cosmological arguments, since it involves the least amount of conceptual baggage and controversial assumptions, is the one given by Samuel Clarke. Like the Kalan Argument it begins with the contingent existential fact that there now exists an aggregate of all the contingent beings there are, but unlike this argument it does not have to invoke any controversial claims about the impossibility of infinite aggregates. It demands an explanation for the existence of this universe on the basis of a more general version of the PSR than the one employed in the Kalan Argument, namely that there is an explanation for the existence of every contingent being, whether or not it ever comes into being. For explanatory purposes the universe itself counts as a contingent being, since it is an aggregate of all the contingent beings there are. It therefore must have a causal explainer. This cause cannot be a contingent being. For if a contingent were to be the cause, it would have to be a cause of every one of the aggregate’s constituents. But since every contingent being is included in this aggregate, it would have to be a cause of itself, which is impossible. The cause, therefore, must be some individual outside the aggregate; and, since an impossible individual cannot cause anything, it must be a necessary being that serves as the causal explainer of the aggregate. This holds whether the aggregate contains a finite or infinite number of contingent beings. Even if there were to be, as is possible for Clarke, an infinite past succession of contingent beings, each causing the existence of its immediate successor, there still would need to be a cause of the entire infinite succession.

It is at this point that David Hume raises what is considered by many to be a decisive objection to Clarke’s argument. He claimed that for any aggregate, whether finite or infinite, if there is for each of its constituents an explanation, there thereby is an explanation for the entire aggregate. Thus, if there were to be an infinite past succession of contingent beings, each of which causally explains the existence of its immediate successor, there would be an explanation for the entire infinite aggregate, and thus no need to go outside it and invoke a necessary being as its cause. Hume’s claim that explanation is in general agglomerative can be shown to be false. For it is possible for there is be a separate explanation for the existence of each constituent in an aggregate, say each part of an automobile, without there thereby being an explanation of the entire aggregate – the automobile. The explanation for the latter would be above and beyond these several separate explanations for the existence of its constituent parts, as for example one that invokes the assembling activity in a Detroit factory.

William Rowe has given a variant version of Clarke’s argument. He chooses as his initial contingent existential fact that there exists at least one contingent being. This is the plaintive cry that one might hear in a coffee shop, “Why is there something rather than nothing?”, to which, according to Sidney Morgenbesser, God’s response is, “Look you guys, suppose I created nothing, you still wouldn’t be happy.” The point of Morgenbesser’s witticism is that even if there were to be nothing, that is, no contingent beings, the PSR still would require that there be an explanation for this big negative fact. The PSR is an equal opportunity explainer, not giving a privileged status to positive reality. We ask “Why is there something rather than nothing?” simply because there happens to be something rather than nothing. The PSR requires there be an explanation for the contingent fact that there exists at least one contingent being. It cannot be given in terms of the causal efficacy of another contingent being, since this would result in a vicious circularity. Thus, it must be in terms of the causal efficacy of a necessary being.

This completes our brief survey of traditional cosmological arguments. It is now time to critically evaluate them. It was seen that each faced an unresolved gap problem consisting in its failure to show that the first cause, unmoved mover, or necessary being has all of the essential divine attributes. The most serious form that the gap problem takes concerns the moral qualities of this being. Herein the problem of evil has been appealed to by the likes of Hume to argue that probably it is not all-good being but rather morally indifferent. As a bumper sticker I once saw had it, “God does exist. He just doesn’t want to get involved.” To counter the challenge of evil it is necessary to construct theodicies for the known evils and give convincing design arguments, which is the topic of the next section.

The most vulnerable premise in these arguments is its PSR, whether in its universal or restricted form. It is imposing on the nontheist opponent of these arguments to ask her to grant that every true contingent proposition (or some restricted set of them) actually has an explanation, for this, in effect, is to grant that the universe is rational through and through. And this occupies almost as high an echelon in one’s wish book as does the existence of God. Hume argued that we can conceive of an uncaused event; and, since whatever is conceivable is possible in reality, PSR is false. Bruce Reichenbach charges that Hume confuses epistemic with ontological conditions. 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 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. What can the defender of the PSR say to get us to give up our prima facie Humean modal intuition? Plainly, the onus is on her, since it is she who uses the PSR as a premise in her cosmological argument.

The cosmological arguer might respond that if the PSR were false, then the universe would be irrational. But it is not rational to believe what is itself irrational. We’ll leave it to the reader to respond to this sophism. Here is another argument for PSR. 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.

Another way of supporting 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 cosmological arguments attempt to establish the cognitive rationality of believing that God exists, they cannot employ a premise that concerns only the pragmatic rationality of believing some proposition, such as the 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.

A more reasonable argument for the PSR is an inductive one based upon our numerous and ever increasing successes in explaining contingently true propositions. The problem with such an inductive argument is that there is a significant difference between the contingent events and objects within the universe which form its inductive sample and the universe as a whole. Thus, it is risky to infer that what holds for the former also holds for the latter.

Recently, we have concocted a new version of Clarke’s cosmological argument that manages to make due with a very weak version of the PSR that requires only that for every contingently true proposition it is possible that it have an explanation, thereby making it more difficult for the argument’s nontheist opponent to reject the PSR premise. Thus it is not required that the proposition reporting the existence of the universe comprised of all the contingent beings there are actually have an explanation, only that it is possible that it does.

Once our opponent has granted the following weak version of the PSR

W-PSR. For every contingently true proposition, p, there is a possible world w that contains the propositions p, q, and that q explains p.

we are able to deduce from it the strong version of the PSR, namely,

S-PSR. For every contingently true proposition, p, there is a proposition q and that q explains p.

in which a possible world is a maximal, compossible conjunction of abstract propositions. It is maximal because for every proposition, p, either p is one of its conjuncts or not-p is; and it is compossible in that all of its conjuncts could be true together. This deduction, which is due to Pruss, goes as follows.

1. For every contingently true proposition, p, there is a possible world w that contains the propositions p, q, and that q explains p. W-PSR

2. p is contingently true and there is no explanation of p. Assumption for indirect proof

3. There is a possible world w that contains the propositions (p and there is no explanation of p), q and that q explains (p and there is no explanation of p). from 1 and 2

4. In w q explains p. true because explanation distributes over a conjunction

5. In w proposition p both does and does not have an explanation.

6. It is not the case that p is contingently true and there is no explanation of p. from 2-5 by indirect proof

7. It is not the case for any proposition p that p is contingently true and there is no explanation of p. from 7

Once we have established by this deduction that there actually is an explanation for the existence of the universe, we show by a series of deductions, which cannot be gone into here, that it is in terms of the free intentional actions of a very intelligent and powerful necessarily existent supernatural being. It must be a necessary being since the universe contains all the contingent beings there are. Since this necessarily existent being freely creates the universe, our argument escapes Schopenhauer’s objection to the cosmological argument as being like a taxicab that we hire and then dismiss when we have reached our destination. For the cosmological arguer begins by demanding, on the basis of the PSR, an explanation for a certain contingent existential fact but when she arrives at our desired destination, God, she dismisses the PSR because she does not require an explanation for the fact that God exists and causes the existence this fact. Since our explainer is a necessary being, it is a self-explaining being in the sense that there is a successful ontological argument for its existence, even if we aren’t smart enough to give it. And, since it freely causes the existence of the universe, it is a self-explaining action for a Libertarian theory of freedom, which is the theory favored by the theist.

Once our opponent realizes that W-PSR logically entails S-PSR, she might no longer grant us W-PSR, charging it with begging the question. Whether an argument begs the question is relative to the epistemic circumstances of its opponent before the argument is given, not after it has been given. But this response would not silence Graham Oppy, for he claims that “once you understand W-PSR properly, you can see that it entails S-PSR; and S-PSR is something which non-theists have good reason to refuse to accept….Those non-theists who were ‘willing to grant W-PSR’ before they heard the argument which Gale and Pruss give should then say that they didn’t fully understand what it was to which they were giving assent.”[1] Herein Oppy is demanding that proper or full understanding be closed under deduction. This demand is completely contrived and has the unwanted consequence that every valid deductive argument, when its premises are fully understood, can rightly be charged with begging the question.

Although Oppy’s demand is unacceptably strong, it still is true that to have an adequate understanding of a proposition one must know some of its entailment relationships. One would not understand, for example, the proposition that this is a material object unless one were prepared to deduce from it that this occupies space. (Please no Castenada-type counter-examples of the “I went to kiss Mary but her lips were not extended” sort!) But, plainly, one can understand that this is a material object without being aware of the very complex propositions that it entails within mereological theory.

We are not able to give a precise criterion for distinguishing between those entailment relations that are constitutive of understanding a given proposition and those that are not, since the concept of understanding is a pragmatic one and thus context-sensitive. But this does not mean that we cannot identify clear-cut cases of someone understanding a proposition and those in which she does not. And certainly one can understand a proposition that uses a modal concept without knowing every theorem of modal logic, just as one can understand a proposition employing geometrical concepts without knowing every theorem of geometry.

The most challenging objection to our argument has been given by Kevin Davey and Rob Clifton. Their strategy is to find a proposition that is strongly incompatible with W-PSR, in that if either is true in any possible world the other is true none, and which is at least as plausible a candidate for being logically possible as is W-PSR. Their candidate for such a proposition is that there is a contingent proposition that lacks an explanation in the actual world, say that there cats or the universe for that matter. This modal intuition seems at first blush to have as much prima facie plausibility as does our modal intuition that every contingent proposition possibly has an explanation. But it turns out that these plausible modal intuitions are strongly incompatible. For W-PSR entails S-PSR and thus that in no possible world is there an unexplained contingent proposition, but the Davey-Clifton intuition entails that there is just such a world.

The strategy that will we adopt for breaking this tie in modal intuitions is to show that one of the two rival modal intuitions coheres better with other of our background modal intuitions. To begin with, our belief in W-PSR coheres better with our proclivity to seek an explanation for any contingently true proposition. That we seek such an explanation shows that we do accept W-PSR, for we would not seek an explanation if we did not believe that it is at least logically possible that there is one. Second, we know what it is like to verify that a given proposition has an explanation, namely by discovering an explanation for it, but we do not know what it is like to verify that a given proposition does not have an explanation: There are just too many possible worlds for that to be accomplished. It is beside the point to respond that we know how to falsify the latter but not the former, since a proposition’s its truth-conditions are directly tied to its conditions of verification, not those for its falsification. These two considerations lend credence to the claim that, in the epistemic order, W-PSR is more deeply entrenched than is the DC claim that it is possible that a given contingent proposition has no explanation. From this conclusion it is reasonable to infer that, in the logical or conceptual order, W-PSR is a better candidate than is the Davey-Clifton proposition for being possible.