09 Nov Is the existence of the universe evidence for the existence of God?
In any version of the cosmological argument, the existence of the universe is the starting point of arguing for the existence of God. Cosmological arguments differ regarding the assumption whether the universe had a beginning or not: The family of Kalam cosmological arguments assume a finite universe, whereas the Leibnizian versions leave the question open. Cosmological arguments can also be classified into deductive arguments and at least partly inductive arguments.
In the first section I will show that all deductive cosmological arguments suffer from a common weakness, which I call the PSR weakness. I will also introduce the notions of P- and C-inductive arguments. In the second section I will briefly examine the evidence for the universe’s having a beginning and finally present my own C-inductive cosmological argument, thereby answering the above question positively.
I. Deductive cosmological arguments: their weakness
I.1 Deductive, P-inductive and C-inductive arguments
An argument is a valid deductive argument if it is inconsistent to suppose that its premises are true but its conclusion false. In other words: In a valid deductive argument, the conclusion follows logically from the premises. The following example is a valid deductive argument:
(Premise 1) All embodied creatures on earth consist of cells.
(Premise 2) My dog is an embodied creature on earth.
(Conclusion) My dog consists of cells.
In contrast, inductive arguments “support” or “give strength” to the conclusion, but do not make it 100 % certain. For example:
(P1) In all hybridizations of homozygous yellow and green peas there is a 75 % probability for the offspring to be yellow.
(P2) Pea plant pp belongs to the offspring of a hybridization of homozygous yellow and green peas.
(C) pp is yellow.
In fact, the probability of pp being yellow is 75 %. Richard Swinburne calls an inductive argument that makes the conclusion probable (i.e. its probability higher than 50 %) a correct P-inductive argument. However, in many cases it is unrealistic to attain a probability of 50 % or higher for the conclusion. We are nevertheless content with such arguments as long as they raise the probability of the conclusion, as in the following case:
(P) In a group of 100 male people sick with cancer, 62 showed a decline in cancer cells after a treatment with curcuma.
(C) Curcuma is remedy for cancer.
Here, we cannot say with certainty that the conclusion is more likely than not, even though 62 % of the study group show the pertaining evidence. Too many questions are open: Do women react in the same way? Is the study group representative? What happened if one did the same study with 100,000 people? Can other factors leading to the decline of cancer cells be excluded? Nevertheless, it would be implausible to deny the conclusion any truth value. The premise certainly makes the conclusion more probable than it would otherwise be. Let’s call, in alignment with Swinburne, such an argument a C-inductive argument.
I.2 The deductive forms of the Leibnizian and Kalam cosmological arguments suffer from a common weakness
I.2.1 The Leibnizian cosmological argument and objections
L1 A contingent being (a being which, if it exists, can not-exist) exists.
L2 This contingent being has a cause or explanation of its existence.
L3 The cause or explanation of its existence is something other than the contingent being itself.
L4 What causes or explains the existence of this contingent being must either be solely other contingent beings or include a non-contingent (necessary) being.
L5 Contingent beings alone cannot cause or explain the existence of a contingent being.
L6 Therefore, what causes or explains the existence of this contingent being must include a non-contingent (necessary) being.
L7 Therefore, a necessary being (a being which, if it exists, cannot not-exist) exists.
Premise L1 is usually taken as the proposition “The universe exists (as a contingent entity)”. This is uncontroversial, just like L3.
Premise L2 may seem self-evident. We have a hard time thinking of things that exist just so, without a cause or explanation. Indeed, this intuition has led to the formulation of the Principle of Sufficient Reason (PSR). The PSR says that it is a priori true that everything has a sufficient reason, cause or ground. But the principle has been contested by many philosophers. First, it seems that it is not an a priori truth. We can just believe that something exists without a reason (as odd as this may seem), without uttering a logical contradiction. Second, deducing from our familiarity with contingent things to which we can ascribe explanations to there being necessarily an explanation for the universe as a whole may not be warranted. Indeed, Bertrand Russell held that deducing the latter from the former is committing the “Fallacy of Composition” (FoC). The FoC consists in concluding from the properties of parts that the whole necessarily has the same properties. For example, a single Lego stone is square, but the building made of such stones can be rectangular. Therefore, so goes the argument, the following argument is not valid:
(P) All parts of the universe are contingent and therefore have a cause.
(C) The universe as a contingent whole has a cause.
I don’t think Russell’s argument succeeds, but that it uncovers an unexpected problem with another premise of the Leibnizian argument. Against the FoC argument I reply thath the concept of “property” used in it is ambiguous. Of course there are many cases in which the whole does not share the properties of its parts; for example, a red tissue does not consist of red atoms. Yet, sometimes the totality necessarily has the same characteristics as its parts. A brick wall is made of bricks; a universe consisting of contingent, caused parts is arguably also contingent and caused. Perhaps the best way to make this clear is as follows: we can imagine every part of the universe to cease to exist; hence every part of the universe is contingent; hence the universe as a whole can cease to exist (for what else would be the case if all parts ceased to exist?) and is therefore contingent.
The problem is that this shows only that a universe composed of contingent parts is itself contingent. It does not show that contingent entities must have a cause. Let us imagine that the universe – that is, the whole of space-time-energy-matter – has always existed, without a beginning; only the arrangements across time differed, making the different physical entities like stars, planets or even creatures contingent. This is prima facie possible. We then would have an uncaused universe with contingent and caused parts. Interestingly, this leaves open the question whether such an uncaused, everlasting universe should be viewed as necessary or contingent. If it could be shown that it is contingent, premise L5 would not hold and we would have another defeater for the Leibnizian cosmological argument.
I.2.2 The Kalam cosmological argument and objections
The Kalam cosmological argument includes the universe’s having a beginning. It can be formulated as follows:
K1 Everything that begins to exist has a cause.
K2 The universe began to exist.
K3 Therefore, the universe has a cause.
Admittedly, premise K2 is debatable. In my view, there are many more good arguments for the truth of K2 than arguments for its only alternative, namely that the universe had no beginning and exists eternally. We shall look at those arguments in section II.
K1 may seem, at first glance, uncontentious. We take it for granted that contingently existing things (to which all the items of our everyday life, except ideas like the concepts of geometric forms or numbers, belong) began to exist at some time and that this coming-into-existence has a cause and therefore an explanation; we often don’t know the cause for something, but to claim that there is no cause at all seems contraintuitive, implausible and somehow mystical, as if one gave up on reason altogether.
Notice that K1 can do with a weaker form of the PSR, namely that everything which has a beginning must have a reason, cause or ground. Let’s call this weaker principle wPSR and consider objections to it and replies to these objections.
I.2.2.1 The modal imagination argument
One can imagine that a brick pops into existence uncaused. Therefore, one might conclude that it is possible that a brick pops into existence uncaused, and hence that the PSR is not a necessary truth.
This popular Humean argument against the PSR can be defeated in several ways.
First, one might argue that in order for this imagination to be a defeater for the PSR, the imagination would have to be possibly actualized in reality. In order to be possibly actualizable, however, it must be actually actualizable. That would mean that there would have to be an actual situation in which something happens without a cause. In how far quantum events count as such will be discussed in section I.2.3. At any rate, it is not clear that the fact that uncaused existence is imaginable makes it a defeater for the necessity of the PSR.
Second, one could doubt that an imagination of things popping into existence uncaused is possible in the first place. Alexander Pruss writes:
…we can imagine a brick coming into existence in the absence of a brickmaker, a brick not resulting from the baking of clay, a brick not made by an angel, demon, or ghost. But that is not the same thing as imagining a brick that comes into existence completely causelessly. To imagine that, we would need to imagine every possible kind of cause – including the unimaginable ones – as absent. That seems to be a feat beyond our abilities.
At any rate, it is not clear that mere imagination warrants valid conclusions concerning the modal status of a proposition. Still, if it were possible to imagine a physical event happening completely uncaused (in the Prussian sense), chances are there might be at least one uncaused event. After all, we can apply criteria of actualizability only to repeatable events we can explore; the universe’s coming-into-existence is by definition a singular event of no kind, hence we cannot be sure that what laws apply to our common physical world also apply to the beginning of the universe.
I.2.2.2 The quantum mechanics objection
A common objection to the PSR uses quantum indeterminacy. If, as it seems, electrons can pass out of existence at one point and come back into existence elsewhere, this opens up the possibility for the universe’s coming into existence without a cause. The state in which the universe was before the Big Bang can only be described in quantum terms and with quantum laws – if it can be described physically at all.
Some theoretical physicists, among them Stephen Hawking, go as far as to contend that the universe came into existence without a cause. According to them, the universe was originally a vacuum with no space-time dimensions. Quantum phenomena which do not obey the PSR made the energy level in the original vacuum state increase rapidly, leading to the Big Bang. Of course, this presupposes that the law of conservation of energy did not apply then.
Two things can be replied here. First, it is still not clear how we should interpret quantum phenomena. There are at least three interpretations, and all of them seem to support the PSR. Second, even the “quantum vacuum origination theory” of the universe is no counterexample to the PSR. A sudden increase in energy, though indeterminate, need not be uncaused. In fact, God’s action is a very good explanation for it (see section III).
I.2.3 Summary: the PSR is the deductive arguments’ main clubfoot
What should have become evident by now is that there are defeaters for the PSR and wPSR respectively on which L2/K1 rest. Though I think that there are much better arguments in favor of the (w)PSR than against it, the principles certainly do not count as a priori, given so much disagreement about it. But to build a deductive argument on premises which are not evident to everyone seems inappropriate. In fact, the strength of the PSR is rather a clubfoot to the argument than a promotor, for it offers ample opportunities for attacks, thus blocking the argumentative chain at step 2. Furthermore, L5 also seems defeasible, as depicted in I.2.3. Thus, even if we could overcome all obstacles concerning the PSR, we would still meet the difficulty of showing that the cause for the universe’s existence must be a necessary one, let alone the theistic God.
So why not leave the deductive path and take the inductive one? Choosing this will also enable us to be open to all kinds of answers to the question “What caused the universe?”, including “Nothing!”. What tips the scale then are the pertaining criteria for assessing explanations, not the requirement of absolute certainty via a deductive argument. Let’s turn to my C-inductive version of the Kalam argument now.
II. Kalam tamed: A new C-inductive version of the KCA
II.1 Scientific and philosophical arguments for a finite universe
A finite universe, i.e. a universe with a temporal beginning, yields better evidence for God’s existence than a temporally infinite universe. The latter does not have a beginning, so the only option for God to come into play here is to be the universe’s sustainer. On a finite account, however, the universe needs not only a sustaining cause, but also a beginning cause. Let’s turn to some arguments for a finite universe.
II.1.1 Scientific arguments
The big bang model is now widely accepted as the standard model for describing the history of the universe. It claims that the universe began with a singularity roughly 15 billion years ago. Arguments supporting that model include the red shift of galaxies and stars, indicating that they move away from each other at an increasing velocity. A hallmark in favor of the big bang theory was reached in 2003 with the Borde-Guth-Vilenkin-theorem which establishes that any universe that has on average over its past history been in a state of cosmic expansion cannot be eternal in the past but must have a space-time boundary, i.e. a beginning. Indeed, according to physicists Barrow and Tipler, this space-time boundary is tantamount to a creation out of nothing: “At this singularity, space and time came into existence; literally nothing existed before the singularity, so, if the Universe originated at such a singularity, we would truly have a creation ex nihilo.”
Another weighty argument for the universe’s being finite is based on the Second Law of Thermodynamics. It states that entropy in a closed system can only increase. Therefore, over an infinite period, a closed system like our universe should have reached the maximum possible entropy, equivalent to a state of equilibrium, which could either be a “hot” one (“hot death” due to crunch of the universe because gravitation wins over expansion) or a “cold” one (“cold death” because the expansion wins over gravitation). Given the fact that we now live in a universe that has neither suffered hot nor cold death, an infinite past of the universe is impossible; hence, the universe must have had a beginning.
II.1.2 Philosophical arguments
William Lane Craig proffers two distinct philosophical arguments against an infinite universe. The first one is aimed at showing that an infinite temporal regress (which a temporally infinite universe requires) is impossible because it is an actual infinite, and an actual infinite cannot exist. An actual infinite is to be differentiated from a potential infinite, which is an “indefinite collection that is at any particular time finite”. Furthermore, though an actual infinite is possible mathematically, it is, according to Craig, impossible metaphysically. He cites the Hilbert Hotel Paradoxes to support this thesis. A Hilbert Hotel is a fully occupied fictional hotel with infinitely many rooms. Suppose, for example, that all but three persons check out. Then, as many persons would have left the hotel as if all guests in rooms with odd numbers had checked out. Other paradoxes can be constructed from that hotel. We can conclude that Hilbert’s hotel is absurd, and hence also any actual infinite cannot exist.
The second argument goes against the possibility of forming an actual infinite by successive addition:
(P1) A collection formed by successive addition cannot be an actual infinite.
(P2) The temporal series of past events is a collection formed by successive addition.
(C) Therefore the temporal series of past events cannot be an actual infinite.
P1 is supported mathematically: The true infinite À0 cannot be reached by successive addition or counting because there is always a number one could add to one’s count, no matter how many numbers one has already counted (hence the impossibility of “counting to infinity”). Furthermore, an actual infinite is an actual À0, not “just” a potential infinite which is at any particular time finite. À0 has no immediate predecessor. There is, therefore, no logical possibility of counting up to it. What about counting down from it? A true infinite cannot be reached either by beginning at an infinitely past point of time and ending now. If a person could count back from infinity and end just now, there would have to be events prior to now in which the person passes a number greater than 0; and before that, an even greater number; and so forth. But when is infinity reached? The only possible answer seems that it is never reached, which would entail that the person never began to count, which is absurd.
From the scientific and philosophical arguments just presented, I conclude that it is highly likely that the universe has a beginning. I will therefore presuppose this in my further argumentation.
II.2 Bayes’s theorem and the best explanation
Bayes’ theorem allows us to calculate a probability for any given hypothesis to be true:
whereby h is the hypothesis, e the evidence and k our background knowledge. P(h/e&k) is the probability of the hypothesis given the evidence and background knowledge. P(e/h&k) captures the probability of the evidence given the hypothesis and background knowledge; in other words, it is the predictive power of h. P(h/k) is the intrinsic probability of the hypothesis, i.e. how likely it is per se given our background knowledge. P(e/k) is the probability of the evidence occurring given only our background knowledge.
When comparing different explanations (hypotheses) for a phenomenon, one can use the Bayes theorem to calculate, at least roughly, the probabilities for the hypotheses and thus compare them. The one with the greatest resulting number wins.
II.3 Why the theistic God is the best explanation for the existence of a finite universe
In order to assess how good theism is as an explanation for the existence of the finite universe, we need to compare it with other available hypotheses. I see in totality three competitors:
- The universe came into existence uncaused.
- The universe was brought about by a number of uncreated god-like beings.
- The universe was brought about by the theistic God.
We already ruled out the option that the universe exists infinitely. The hypothesis of one or more created beings bringing about the universe only shifts the explanandum back in time and is really a version either of theism or (2).
The probability of (1) is low. P(e/h&k) is low on this account, because we simply don’t expect anything to happen uncaused (cf. the PSR). P(h/k) fares no better; one the one hand, it is an extremely simple hypothesis, but on the other hand, it does not fit at all with our background knowledge. P(e/k) is hard to assess; how can we estimate the probability of the existence of the universe given our background knowledge, when our background knowledge derives mainly from that very universe? We could, however, follow our intuition and marvel at there being anything at all. It is a reasonable principle to assume that the most “natural” state of things is that there is simply nothing, for every thing that exists requires some effort. Given that principle, P(e/k) will be low. It will stay the same throughout all the examined hypotheses; therefore, we can focus on the other two terms. All in all, P(h/e&k) of (1) is low.
What about (2)? P(e/h&k) is higher than with (1), because assuming a cause makes the occurrence of a universe much more likely and is also better in harmony with our background knowledge. P(h/k) of (2) is somewhat ambivalent: On the one hand, it goes better with our background knowledge, one the other hand it is a more complex hypothesis than (1). So let us assume the two probabilities here to be on a par. The overall probability of (2) is hence higher than that of (1)
The theistic hypothesis (3) fares best of all. That is, first, because P(e/h&k) is at least as high as with (2); in both cases, there are personal agents bringing about the universe. However, P(h/k)is higher on (3) than on (2), because (3)’s simplicity is greater than (2)’s; clearly, it is a simpler hypothesis to postulate one God than many gods. Thus, the overall probability of (3) is highest of all three competing hypotheses.
We’ve seen that deductive cosmological arguments are a problematic battleground because they require strong metaphysical and logical principles that can be defeated. Inductive reasoning allows us to select the best explanation; if we compare all the feasible explanations, the one winning the day is the one we should adopt. When presupposing a finite universe, as one should given the strong scientific and philosophical support, the only three feasible hypotheses are that the universe came into existence uncaused, that it was created by a multitude of god-like beings or that it was created by the theistic God. The latter hypothesis clearly proves the most likely one. The universe is good evidence for the existence of God, and God is the best explanation for the existence of the universe.
 Swinburne 2004, p. 6
 Swinburne 2004, p. 6
 Reichenbach 2004, p. 104
 Reichenbach takes this outline to apply to all cosmological arguments. However, the Kalam version has a more specific form. A look at Leibniz’s On the Ultimate Origination of Things makes it clear that the presented form is that of a Leibnizian argument.
 This outline is taken from Craig & Sinclair 2009, p. 102.
 Pruss 2009, p. 48
 Namely (1) the Copenhagen interpretation, according to which the observer brings about some events that seem uncaused; (2) the Bohmian view that the seeming indeterminacy is due to hidden variables (3) the Many Worlds Interpretation, which postulates a myriad of parallel worlds in which the different possible outcomes of quantum experiments obtain.
 Barrow & Tipler 1985, p. 442
 Craig 2013
 Craig 2013, p.8
 Craig 2013, p.11 (adapted)
Barrow, John and Tipler, Frank (1985): The Anthropic Cosmological Principle. Oxford: Clarendon Press
Craig, William Lane (2013): The Kalam argument. In: Moreland, J.P. & Sweis, Khaldoun A. (ed.): Debating Christian Theism, Oxford University Press, pp. 7-18
Craig, William Lane and Sinclair, James D. (2009): The kalam cosmological argument. In: Craig, W.L. & Moreland, J.P.: The Blackwell Companion to Natural Theology, Wiley-Blackwell, pp. 101-201
Pruss, Alexander (2009): The Leibnizian Cosmological Argument. In: Craig, W.L. & Moreland, J.P.: The Blackwell Companion to Natural Theology, Wiley-Blackwell, pp. 24-98
Reichenbach, Bruce (2004): Explanation and the Cosmological Argument. In: Peterson, Michael & VanArragon, Raymond (ed.): Contemporary Debates in Philosophy of Religion, Blackwell
Swinburne, Richard (2004): The Existence of God (2nd edition). Oxford University Press