Some Comments on William Craig's "Creation and Big Bang Cosmology"

Adolf Grünbaum


University of Pittsburgh


[This article was originally published in Philosophia Naturalis, Vol. 31, No. 2, pp. 225-236, 1994.]

I. Introduction

Craig makes a whole array of assertions, some of which call for much more detailed scrutiny than is possible in a mere commentary. Elsewhere, I shall publish a detailed critique of his thesis that the kalam cosmological argument justifies a creationist interpretation of the Big Bang world (Craig, 1992, pp. 237-240) as well as a full reply to his scurrilous attack on my views (Craig, 1992). Indeed, he offers metaphysical a priori considerations independently of modern cosmology as his centerpiece to argue that the universe has a creative cause of its existence (1979, Part II; 1992, pp. 233-236). But he relies on scientific cosmology as further support for claiming that the universe began to exist. In so doing, he disingenuously makes much of a relativistically impermissible Big Bang model to argue that "currently accepted cosmological theory does lend tangible support to the theistic doctrine of creatio ex nihilo." A crucial premise in his argument is avowedly "the metaphysical intuition that something cannot come out of absolutely nothing." More explicitly, Craig takes the following proposition (1) be metaphysically necessary:

(1) Anything that begins to exist but does not have a transformative cause must have a creative cause ex nihilo, rather than no cause at all.

Craig deploys this modal proposition (1) to argue for the divine creation ex nihilo of the Big Bang universe. Hence I need to comment on the physics of that universe to show first how Craig misappropriates it in the service of his cause.

II. The Big Bang universe of the general theory of relativity

A. The Big Bang in the Event-Ontology of the General Theory of Relativity ("GTR").

In my earlier writings (Grünbaum, 1989; 1990a; 1990b; 1991), I had discussed two Big Bang models, which I called Case (i) and Case (ii), respectively, and which I am about to characterize. Yet, as I noted then, Case (i) is not a bona fide model of the GTR for reasons given in the event-ontology of that theory. In the Case (i) model, the Big Bang is supposedly the temporally first physical event of the space-time, and is said to occur at the instant t = 0. But, as we shall now see, the Big Bang does not meet the requirements for being a bona fide physical event in the GTR. Instead, the Big Bang is a so-called "singularity" in at least the sense that, as we approach it from ever earlier moments, the space-time metric of the GTR becomes degenerate, and the scalar curvature as well as the density become infinite (Wald, 1984, pp. 99-100).

John Stachel (1993) has justified the view that this singular status robs the Big Bang of its event-status in the GTR. As he showed, points of the theoretical manifold first acquire the physical significance of being events, when they stand in the chrono-geometric relations specified by the space-time metric, which does double duty as the gravitational field in the GTR.

Thus, in the GTR, it turns out that "the notion of an event makes physical sense only when [both] manifold and metric structures are [well] defined around it" (Wald, 1984, p. 213; cf. also Hawking & Ellis, 1973, p. 56). And in that theory, space-time is taken to be "the collection of all [physical] events" (Hawking & Ellis, 1973, p. 56). Thus, the Big Bang does NOT qualify as a physical point-event of the space-time to which one would assign three spatial coordinates, and one time coordinate. Therefore, contrary to Case (i), the past cosmic time-interval is open or unbounded at t = 0, rather than closed or bounded by a first moment. Despite the ontological illegitimacy of the Case (i) model, I discussed it in earlier writings because Pope Pius XII, Sir Bernard Lovell, and William Craig each claimed support from it for divine creation ex nihilo. Besides, the Case (i) model had figured in the astrophysicist Narlikar's secular creationism (Grünbaum, 1993).

But, as we saw, the Big Bang singularity t = 0 is actually excluded as not being a physical event occurring at an actual moment of time. Thus construed, the relativistically bona fide Big Bang models differ from those in Case (i) by being temporally unbounded (open) in the past. And hence the past physical career of the Big Bang universe did NOT include a first physical event or state at which it could be said to have begun. I designated the bona fide models as Case (ii) models.

However in either Case (i) or Case (ii), the current age of the Big Bang universe is metrically of finite duration, say 15 billion years, depending numerically on the time-rate of its expansion (Wald, 1984, p. 99). Importantly, there are good reasons in the GTR for claiming that no instants of time whatever existed before that finite time-interval in either Case (i) or Case (ii) (Wald, p. 99). Thus, even if the singular Big Bang were included as having occurred at a bona fide moment of time t = 0, this hypothetical instant had no temporal predecessor.

As we now see, physical processes of some sort already existed at every actual instant of past time. After all, despite the finite duration of the past, there was no time at all at which the physical world did not exist yet. Thus, we can say that the Big Bang universe always existed, although its age is only 15 billion years. Here, the word "always" means "for all actual times," but it does not guarantee that time, past or future, is of infinite duration.

As we saw, in the Case (i) world, there did not exist any instants of cosmic time before t = 0. Therefore, no supposed EARLIER cause could possibly have been operative before t = 0. For that reason alone, the Big Bang at the putative initial instant could not have had any temporally prior cause. A fortiori, it could not have been temporally preceded by a fictitious state of utter nothingness, whatever that is.

Let me take for granted the altogether reasonable view that only events can qualify as the momentary effects of other events, or of the action of an agency. Since the Big Bang singularity is technically a non-event, and t = 0 is not a bona fide time of its occurrence, the singularity cannot be the effect of any cause in the case of either event-causation or agent-causation alike. By the same token, the singularity t = 0 cannot have a cause, either earlier or simultaneous, be it natural or supernatural! It is crucial that, in particular, it cannot have a simultaneous cause.

Yet Craig insists that since the singularity admittedly had no earlier cause, it must have had a simultaneous one, because it is metaphysically impossible that it be uncaused or "come out of absolutely nothing." And he charges me with having overlooked this "obvious alternative" of a simultaneous cause, telling us in an incoherent and exasperatingly obscure footnote that the Big Bang singularity and its cause "both occur coincidentally (in the literal sense of the word) that is, they both occur at t0" (footnote 1). But besides being vitiated by illicitly treating the singularity as a bona fide effect requiring a cause, and a simultaneous cause in particular, Craig's case suffers from an array of other difficulties.

1. Craig calls for a supernatural simultaneous cause, by claiming that, in its absence, the Big Bang would be coming out of absolutely nothing, which he deems impossible. But in what sense could an uncaused Big Bang universe be thought to have "come out of absolutely nothing"? Surely NOT in the sense that there existed moments of time before the Big Bang at which the physical universe did not yet exist but only nothing. As I argued above, despite the metrically finite past duration of the Big Bang world, there was no such prior time. Thus, the finitude of that past does not warrant the conclusion that if this universe is uncaused, it must have "come out of" a prior state of nothing.

Craig assures us (footnote 4) that he is not supposing the existence of "empty time prior to the singularity," when he insists on his "metaphysical intuition" that "something cannot come out of absolutely nothing." Instead, he means to assert: "it is false that something existed prior to the singularity." The latter assertion is correct. Thus, Craig's claim reduces to asserting that a universe whose past is finite cannot exist uncaused. But he has no objection to the finitude of the past. On the contrary, he insists that "an infinite temporal regress of events cannot exist" (1979, p. 99).

Immediately after affirming his stated "metaphysical intuition," he writes:

A pure potentiality cannot actualize itself.... On the theistic hypothesis, the potentiality of the universe's existence lay in the power of God to create it. On the atheistic hypothesis, there did not even exist the potentiality for the existence of the universe. But then it seems inconceivable that the universe should become actual if there did not exist any potentiality for its existence. It seems to me therefore that a little reflection leads us to the conclusion that the origin of the universe had a cause.

Here Craig is telling us that an external cause is required to effect the realization of "the [mere] potentiality of the universe's existence," and that if the latter potentiality did not exist, "then it seems inconceivable that the universe should become actual" (italics added).

But in what sense and on what grounds does Craig claim that the universe did become actual if, as he grants, there was no "empty time prior to the singularity"? It does not follow from the de facto or actual existence of the universe that it ever "became" actual! Precisely by postulating such "becoming actual," Craig begs the question when offers such purported actualization as grounds for inferring that there must have been a "potentiality of the universe's existence." Thereupon, he infers that the realization of that potentiality required an external cause, which he identifies with divine creation. Thus, Craig's causality is a phantom.

But as I shall now explain, another error, both inveterate and subtle, has lent spurious plausibility to the potentiality on which Craig relies to lend credence to his stated metaphysical intuition.

2. The notion that an uncaused universe is impossible tout court, which Craig finds inescapable, is rooted in a blunder that has a longer philosophic history. Thomas Aquinas, Descartes, Berkeley, Leibniz and Locke, as well as the contemporary theists like Richard Swinburne, John Polkinghorne and Philip Quinn have told us that the world must have a ratio essendi, which supplies an answer to the question "Why is there anything at all rather than just nothing?"

But I claim that the question "Why is there anything at all rather than just nothing?" is a misguided query, at least to the extent that it calls for a cause external to the universe. Thus, it is wrong-headed, I shall now contend, to ask for the external cause or reason of the bare existence and persistence of the world, its so-called ratio essendi. But it is vital to distinguish such a supposed cause, as Aquinas did, from a merely transformative cause, which just produces changes in things that already exist, or generates new entities from previously existing objects.

There is a crucial underlying assumption that animates the theological creationist and conservationist ratio essendi given by a galaxy of theists. Oddly enough, they take it to be axiomatic that if there is a physical world at all, then its spontaneous, undisturbed or natural state is one of utter nothingness, whatever that is. Those many theists who make this strange assumption, have thereby generated grounds for claiming that the very existence of matter, energy or whatever constitutes a deviation from the alleged spontaneity of nothingness. And that supposed deviation must then have a suitably potent external cause. Aquinas claimed to have established this existential causal dependence of the world on God -- or creation ex nihilo -- on philosophical grounds, whereas he saw the belief in the initial temporal creation of the world to be a matter of Christian faith.

Just this assumption of spontaneous nothingness is at least insinuated by the biblical story of Genesis. But Aquinas, among others, makes it explicit. He used the loaded word "creature" to refer to any contingent entity and declared: "the being of every creature depends on God, so that not for a moment could it subsist, but would fall into nothingness were it not kept in being by the operation of Divine power" (quoted in Quinn, 1993, p. 593; my emphasis). Thus, here we have the crucial presupposition: There would be just nothingness, whatever that is, were it not for divine creative or conservative intervention (Grünbaum, 1995).

But what, oh what, I must ask, is the evidence for this philosophically fateful assumption of the spontaneity of nothingness? Why, in the absence of an external supernatural cause, should there be just nothing, even if we are clear just what that would mean? I am rather at a loss to understand why a galaxy of philosophers thought that the mere logical or empirical contingency of the existence of any given particulars can support the spontaneity of utter nothingness. It appears that the theological presupposition of the spontaneity of nothingness lacks even the most rudimentary plausibility. It is a mere figment of the imagination or, as one might say in German, "ein Hirngespinst." Moreover, some philosophers, such as Henri Bergson, have asserted the unintelligibility of the notion of absolute Nothingness.

The tacit presupposition of spontaneous nothingness also contributed to Leibniz's demand for a necessary being to provide a sufficient reason for the existence and persistence of contingent things.

As I have just argued, the seminal question as to the ratio essendi of the world of contingent beings, far from being innocent and imperative, has forfeited the rationale that animated it at the hands of such major figures as Aquinas and Descartes. Their problem turns out to have been a pseudo-problem. And their proposed theological resolution of it is a pseudo-explanation. One cannot overestimate, I believe, the extent to which the dubious rationale for a ratio essendi unconsciously insinuates itself to confer spurious plausibility on that pseudo-explanation.

3. We recall that Craig (1992, and 1979, Part II), affirms the following modal proposition (1) in the context of the Case (i) and Case (ii) Big Bang models: Anything that begins to exist at an initial instant or has a metrically finite yet unbounded past, but does not have a transformative cause, must have a creative cause ex nihilo, rather than no cause at all. To this, I have a two-fold response: (i) A priori one can assert only the logical possibility that an entity which begins to exist (in one of the two specified senses) has either a transformative cause or a creative cause ex nihilo or no (efficient) cause at all, and (ii) There is no empirical evidence to support even a non-modal, logically contingent version of Craig's proposition (1).

As for my response (i), the theist Richard Swinburne (1968, p. 305) rightly contends that the mere failure of a cosmological theory to provide a cause for a putative initial state of the universe cannot be taken to count against that theory, because it is not a necessary truth that there be such a cause. Thus, Swinburne rejects Craig's modal premise (1) as false, claiming, in effect, that if it is indeed true that anything which begins to exist has an efficient cause of its existence, then it is only contingently true. By the same token, Quentin Smith (1993, ch. VI, Section 4, pp. 178-185) cogently refutes Craig's various arguments for his modal kalam proposition (1) and indeed shows it to be false by special reference to the Big Bang universe.

As for my response (ii) that there is no empirical evidence for a non-modal, logically contingent version of Craig's proposition (1), it would be unavailing to his theological thesis if the original 1948 version of the steady-state cosmology (due to Bondi and Gold, and to Hoyle) had turned out to be well-supported empirically. That theory featured the "Perfect Cosmological Principle," which asserts the conservation of the density of matter in a temporally infinite universe. Clearly, in the steady-state world, the expansion of the universe amid nomic density-conservation is the creative cause of the popping into existence of new matter ex nihilo. Thus, the steady-state theory provides a physical, rather than supernatural, creative cause for the coming into being of its new matter. Evidently, even if the original steady-state theory had turned out to have good observational credentials, it would not warrant that the universe as a whole had a creative cause, let alone that it was an external cause identifiable with God, as required by Craig's scenario.

4. I argued initially that the Big Bang cannot have a simultaneous cause, let alone a divine one, because it does not qualify as an event, and a fortiori cannot be the effect of any temporally prior or simultaneous cause, either natural or supernatural. Having accused me of having failed to consider the "obvious alternative" that the cause of the Big Bang was simultaneous with it, Craig explains that "The [supernatural] cause of the origin of the universe is causally prior to the Big Bang, though not temporally prior to the Big Bang." Therefore let me comment on his contentions concerning asymmetric instantaneous causation in natural contexts, and supposed instantaneous divine asymmetric causation of the Big Bang.

The proponent of simultaneous asymmetric causation must give us a criterion for distinguishing one of two causally connected simultaneous events as the cause of the other. Clearly, for simultaneous events, temporal priority is unavailable to provide the required criterion for causal priority. Speaking of "philosophical discussions of causal directionality," Craig tells us that they "routinely treat simultaneous causation, the question being how to distinguish A as the cause and B as the effect when these occur together at the same time." He then cites five articles by six authors. But, alas, nary a word from Craig as to just what criterion of causal priority that dispenses with temporal priority he is invoking from the literature, and why we should consider it viable.

James Woodward, who is a student of that literature, has reported (private communication, April 14, 1994), that "there is no generally accepted account of the direction of causation and a fortiori no generally accepted account that makes causal order independent of time order." And he has expressed the view (private communication, January 7, 1994) that there is no other account of causal direction besides Daniel Hausman's (1984; 1993) which does not rely on time order and even begins to work.

Woodward (private communication, January 7, 1994) then considers the question whether one could appeal to Hausman's account of causal priority to make sense of the theist's claim that God's volition (or whatever) was causally prior to the Big Bang or to some other event at which the universe originated. And after stating Hausman's criterion of causal priority, Woodward instructively replies concerning theistic creation as follows:

If there is a well-defined sense in which God's choice or volition (V) to create is causally prior to the origination event O, then [on Hausman's criterion] it should be the case that anything that is causally connected to V is also causally connected to O, but there is at least one thing causally connected to O which is causally independent of V. Alternatively, formulated in a more hypothetical mode, the idea would be that any intervention which changes V should also change O, but it should be possible in principle to wiggle O in a way that is independent of V. Put this way, it seems to me that there are two features of the application of this account to the creation of the universe which I would have thought that theists would want to resist. First of all, I take it that theists don't think that anything causes God's volitions V -- these are supposed to be the expression of God's free will and there is no well-defined sense in which anyone or anything can manipulate these and see whether there are corresponding changes in O. Second, I would have thought that it was part of the logic of the theist's argument that the only event which could possibly causally influence O is V -- that there is no other path by which O could be caused or causally manipulated except via V. God's will is supposed to be a sufficient cause for -- and the only possible cause of -- the origin of the universe. So both clauses in Hausman's account will fail to apply in this case, and, at least within the framework of that account, there will be no well-defined sense in which V is causally prior to O.
... So I think you can legitimately conclude that, at least as far as the extant accounts of causal direction go, the theist is unable to explain in what sense V is causally prior to O.

Clearly, Craig has not provided an explication of that sense, let alone for instantaneous divine creation, by telling us (footnote 2):

In the case of God's creating the universe, it is, of course, evident which is the cause and which the effect, since it is metaphysically impossible for God to have an external cause.

In any case, in the face of my demonstration above of the failure of Craig's argument for a simultaneous cause of the Big Bang, this Craigean declaration is unavailing. But he also tells us that there are mundane cases of instantaneous causation that qualify as asymmetric, one-way or directed. And he cites Kant's example of a ball denting a cushion. Another example is supposedly furnished by a locomotive pulling a train.

Clearly, Kant's example is predicated on the assumption that, prior to the ball's impact on the cushion, the dent was not present in the cushion. Hence we must consider the process which issued in that dent. For the moment, let us ignore the principle enunciated by the special theory of relativity that the finite speed of light places an upper bound on the transmission of causal influences, and restrict ourselves to an analysis based on Newtonian physics in the relevant contect of event-causation. Then it seems clear that there was an (ideally elastic) interaction of momentum-exchange between the ball and the cushion as follows: The cushion reacts to being dented by the ball so as to change the ball's momentum instantaneously. In other words, just as the ball transfers some of its initial momentum to a portion of the cushion, the cushion reacts by intercepting the motion of the ball. Thus, in Kant's ball-and-cushion example, the dynamic causation is symmetric rather than asymmetric, even if instantaneous. By the same token, there is symmetric causation in processes governed by the conservation of linear momentum, such as the collision of two billiard balls.

In the case of the locomotive, it is false to say that the locomotive is tugging at the train cars but not conversely; on the contrary, by Newton's third law of instantaneous action and reaction, each of them exerts the same amount of force on the other (though in opposite directions). Yet the locomotive is subject not only to the reactive force of the cars that would hold it back, but also to the forward reactive force of the ground on its wheels. Thus, the locomotive, and thereby the cars, experience the same net forward acceleration.

It is a commonplace that, if the special theory of relativity is taken into account, a force applied to one end of a solid rod lying on a table will not affect the other end simultaneously, because the transmission of an influence cannot be instantaneous in that theory. Similarly for Kant's cushion-example and the case of the locomotive. In a footnote, Craig dismisses this denial of instantaneous causation. I regret to say that I can discern only question-begging babble in what he says there.

Worse, for good measure, Craig adds that, according to some philosophers, "all efficient causation is simultaneous." It is the old maxim "Cessante causa, cessat effectus." The argument given for this claim as stated by Craig is a patent non-sequitur and yields a false conclusion, thereby illustrating the pitfalls of such a priori theorizing about causation. The empirical absurdity of the claim is illustrated by the mundane fact that a woman does not cease to be pregnant at the moment she ceases to have the sexual intercourse that impregnated her. Medicine furnishes other counter-examples: In carcinogenesis, the activation of an oncogen far antedates the formation of a malignancy. And exposure to the pathogen of an infection antedates the onset of the infection. Indeed, examples of time-delayed causation are, of course, legion.

The so-called "laws of coexistence" relate simultaneous quantities such that they are causally connected to one another. Examples of such laws are Newton's second and third laws of motion, his law of universal gravitation, Einstein's E=mc2, the Stefan-Boltzmann radiation law, the gas laws of Boyle, Gay-Lussac, and van der Waals, as well as Ohm's law. It would be interesting if some viable criterion of causal priority, as applied to any one such law, could single out one of the simultaneous physical quantities as the asymmetric cause of the other.

Suppose we may assume that some explication of causal priority is conceptually sound and licenses the empirical possibility of asymmetric simultaneous causation. Then the question still remains what bona fide examples of such causation actually exist in the world. Craig has provided no cogent reason to conclude that there are any. I do not deny that there are, but I need to be shown.

Finally, as Craig notes, I have argued (Grünbaum, 1990) that Lévy-Leblond was misguided in feeling a need to obviate the finite age of the universe (in, say, years) by invoking Charles Misner's non-standard time-metrization, which confers an infinite duration on the past. Since I have been concerned to examine Craig's argument for divine creatio ex nihilo. I need not comment on the merits of his objections to Misner's alternative time-metrization as such. I have contended that Craig's metaphysical quest for a simultaneous cause of the Big Bang is generated by a pseudo-problem and that his proposed theological solution of it is without merit. Hence I likewise deny the cogency of his complaint that Misner's metrization fails to solve the pseudo-problem posed by Craig's "metaphysical intuition," which calls for the necessity of a ratio essendi of the Big Bang world.


Craig, W.L. (1979). The Kalam Cosmological Argument, Harper & Row, New York.

Craig, W.L. (1992). The Origin and Creation of the Universe: A Reply to Adolf Grünbaum, British Journal for the Philosophy of Science 43, 233-240.

Grünbaum, A. (1989). The Pseudo-Problem of Creation in Physical Cosmology. Philosophy of Science 56, 373-394.

Grünbaum, A. (1990a). The Pseudo-Problem of Creation in Physical Cosmology. In: Leslie J. (ed.), Physical Cosmology & Philosophy. New York: Macmillan, pp. 92-112. This is a reprint of Grünbaum (1989). The paper was also reprinted in Epistemologia (1989), 12, 3-32, and in Free Inquiry (1990), 9, 48-57.
A German translation, Die Schöpfung als Scheinproblem der physikalischen Kosmologie, appeared in: Bohnen, A. & Musgrave A. (eds.) (1991), Wege der Vernunft, Festschrift for Hans Albert. Tübingen: J.C.B. Mohr (Paul Siebeck), pp. 164-191.

Grünbaum, A. (1990b). Pseudo-Creation of the Big Bang. Nature 344, 821-822.

Grünbaum, A. (1991). Creation as a Pseudo-Explanation in Current Physical Cosmology. Erkenntnis 35 , 233-254.

Grünbaum, A. (1993). Narlikar's "Creation" of the Big Bang Universe Was a Mere Origination. Philosophy of Science 60, 638-646.

Grünbaum, A. (1995). Origin versus Creation in Physical Cosmology. To appear in Falkenburg, B. & Krüger, L. (eds.), Physik, Philosophie und die Einheit der Wissenschaft. Mannheim: B.I. Wissenschaftsverlag.

Hausman, D. (1984). Causal Priority. Noûs 18, 261-279.

Hausman, D. (1993). Linking Causal and Explanatory Asymmetry. Philosophy of Science 60, 335-451.

Hawking, S. and Ellis, G. (1973). The Large-Scale Structure of Space-Time. New York: Cambridge University Press.

Quinn, P. (1993). Creation, Conservation, and the Big Bang. In: Earman, J. et al. (eds.), Philosophical Problems of the Internal and External Worlds: Essays Concerning the Philosophy of Adolf Grünbaum. Pittsburgh and Konstanz: University of Pittsburgh Press and Universitätsverlag Konstanz.

Smith, Q. (1993). A Criticism of A Posteriori and A Priori Arguments for a Cause of the Big Bang Singularity. In: Craig, W.L. and Smith, Q., Theism, Atheism, and Big Bang Cosmology. Oxford: Clarendon Press.

Stachel, J. (1993). The Meaning of General Covariance. In: Earman, J. et al. (eds.), Philosophical Problems of the Internal and External Worlds: Essays Concerning the Philosophy of Adolf Grünbaum. Pittsburgh and Konstanz: University of Pittsburgh Press and Universitätsverlag Konstanz.

Swinburne, R. (1968). Space and Time. New York: St. Martin's Press; London: Macmillan.

Wald, R. (1984). General Relativity. Chicago: University of Chicago Press.

"Some Comments on William Craig's 'Creation and Big Bang Cosmology'" is copyright © 1994 by Adolf Grünbaum. The electronic version is reprinted with the permission of Philosophia Naturalis and copyright © 1999 Internet Infidels. All Rights Reserved.