Analysis of William Lane Craig versus Sean Carroll February 2014 Debate

Philosopher William Lane Craig and Physicist Sean Carroll debated the subject of God and Cosmology on February 21 in New Orleans.  Out of all the debates of Dr. Craig's that I have seen (which are many), this one was definitely one of the best.  Both debaters carried themselves professionally, made a reasonable effort to respond to each other's arguments, and managed to keep the debate both informative and entertaining.

I will confess at the outset that I generally find Dr. Craig's arguments to be sharp and persuasive, though I do not sympathize with his Christianity.  Although I do end up mostly defending Dr. Craig's position and criticizing Dr. Carroll's position in this analysis, I attempted to do so fairly and with attention to arguments.  I believe this is in alignment with the spirit of debate and also makes for an enjoyable exercise of philosophy.

Given the debate structure and my interests in the topics discussed, I've divided this analysis into three main parts: Dr. Carroll's positive case, Dr. Craig's Positive Case, and Dr. Carroll's Negative Case.  The positive cases comprise each debater's arguments in favor of their preferred world views: naturalism and theism, respectively.  Those who are familiar with Dr. Craig's debates may wish to skip my discussion of his positive case because they are likely familiar with it.  Dr. Carroll's Negative Case comprises his arguments against Craig's theistic arguments.  I finish the article with a few concluding remarks.

Dr. Carroll's Positive Case

For his positive thesis, Dr. Carroll advanced naturalism as the best world view in light of cosmological considerations.  He said this in his opening speech:

"Naturalism says that all that exists is one world--the natural world--obeying laws of nature which science can help us discover. Theism says that in addition to the natural world there is something else, at the very least God, perhaps there are other things as well. I want to argue that naturalism is far and away the winner when it comes to cosmological explanation."

Dr. Carroll went on to list three primary points in support of his position:

1. Naturalism works
2. The evidence is against theism
3. Theism is not well-defined

The only of these three points that lends itself to a positive justification of naturalism is (1).  Numbers (2) and (3) are part of Carroll's "negative" case against theism.  I will consider (1) now and the latter two points in a section below.

After several viewings of Dr. Carroll's speeches, I am not sure I saw an explicit defense of (1) throughout his talks.  He primarily argued against theism rather than for naturalism, so (1) was not the main thrust of his position.  Nevertheless, he did list (1) as an objective of his, and he did propose naturalism as his alternative to theism, so point (1) deserves some attention.  I will attempt to piece together how Carroll might have meant (1) to be understood from what he said throughout the debate and then give my own analysis of it. (Dr. Craig did not himself seem to address (1) directly either.)

Carroll might argue in defense of (1) that, for any observed phenomenon in the cosmos, it is possible to build a naturalistic model of the phenomenon that, though not necessarily correct, makes predictions and furthers our understanding of the phenomenon.  He reiterated often throughout the debate that many (if not all) naturalistic models of the fundamental nature of the cosmos are problematic.  Thus, he was careful to point out that he was not saying this or that model is definitive or correct; rather, he seemed to be suggesting that, in the long run, if scientists keep building and testing naturalistic models of the universe, these models will converge on the truth.  I suppose it could be in this sense of "eventual convergence on the truth" that he takes naturalism to "work."  Since he was not offering any definitive model of the cosmos as "The One True Naturalist Model," he must not take the proposition "naturalism works" to mean "naturalism is true because this naturalistic model is true."  Instead, his position appeared to be that, although we do not have The One True Naturalist Model in hand yet, we know enough from our existing models to know that, whatever model of the cosmos is ultimately correct, it will be a naturalistic model.

Now, if this is a fairly good approximation of what Carroll meant, it seems to me that Carroll's defense of (1) would be ineffectively stipulative.  In the absence of The One True Naturalist Model to demonstrate the efficacy of naturalist metaphysics, it is difficult for me to see how a collection of non-definitive naturalist models amounts to a demonstration of the adequacy of naturalistic explanation.  If these models are incorrect or incomplete (as Carroll candidly admitted), how do these models in any way show or demonstrate that the actual world is a naturalistic one?  Analogously, if scientists were to conjure a bunch of models of an Earth-centered solar system, how would the existence of such models support the claim that Earth is in fact the center of the solar system?  It seems clear to me that the mere existence of models sharing some common feature X do not amount to any defense of the claim "X is true."  For that claim, one would need a correct model rather than a host of incorrect ones.

Alternatively, Carroll might take the claim "naturalism works" to mean that the physical sciences have made a lot of progress by solely relying on an ontology of only physical forces and objects.  While it seems true that such progress has been made by the sciences in certain domains, it does not follow from this that such progress will be made by the sciences universally across all domains of inquiry.  Successfully building computers or jet engines under naturalistic assumptions, for example, tells us nothing about whether God exists, immaterial human minds exist, or free will exists.  Edward Feser has offered a useful analogy for grasping this point.¹ He suggests that a method of science can be likened to a metal detector.  While it's true that a metal detector is very good at detecting metallic objects, it is no good at detecting anything else.  However, given the fact that the metal detector is really good at detecting metallic objects and only metallic objects, it in no way follows from this that only metallic objects exist.  The success of naturalistic science is in a similar situation.  Although the sciences are really good at detecting certain truths about the physical world, it does not follow from this that only the physical world exists.  Thus, even if we grant Sean Carroll that naturalism "works" in the sense that it's made progress in certain domains, it does not amount to a vindication of a universal naturalist ontology.  As such, Dr. Carroll's assertion that "naturalism works" cannot be treated synonymously with the assertion "naturalism is true."  It remains a bit ambiguous to me from what Carroll said in the debate whether he sought to establish that naturalism were true, but, to the extent that he did seek to do that, I cannot see how he would have been successful in doing so.

In sum, then, I think Dr. Carroll's positive case for naturalism is either ineffectively stipulative (if we take his claim "naturalism works" to mean "naturalist models will eventually converge on the truth"), or incomplete (if he takes the claim "naturalism works" to mean "naturalism is true").

(Note: At the end of Dr. Carroll's first speech, he did provide a list of features of the world that he thought made more sense given naturalism than they did given theism, features like the imperfection of the world, the dis-uniformity of moral law across cultures and times, the existence of suffering, etc.  He could have meant for this to amount to a defense of the claim that naturalism works; but, given the breadth of topics raised there, the complexity of each of the topics, and the brevity with which he discussed them, it does not seem practical to attempt to discuss them here.  I will thus just let this portion of his speech dangle for the reader to consider on his or her own.)

Dr. Craig's Positive Case

As the affirmative side in the debate, Dr. Craig offered the following positive case in his opening speech:

"The evidence of contemporary cosmology actually renders God's existence considerably more probable than it would have been without it….Contemporary cosmology provides significant evidence in support of premises in philosophical arguments for conclusions having theological significance."

Dr. Craig was careful to note that he was not arguing that one could go immediately and directly from some cosmological evidence to the conclusion that God exists.  Instead, Dr. Craig wanted to argue that some cosmological evidence supported the plausibility of a premise in a philosophical argument for God's existence.  He went on to offer two familiar arguments to this effect: the Kalam Cosmological Argument and the Fine-Tuning Argument.

The Kalam Cosmological Argument

Craig's version of the Kalam argument ran as follows:

1. If the universe began to exist, then there is a transcendent cause which brought the universe into existence.
2. The universe began to exist.
3. Therefore, there is a transcendent cause which brought the universe into being.

Craig took premise (1) to be uncontroversial (though, to his surprise, Sean Carroll did challenge it in the debate) so he primarily argued for (2) in his opening speech.  Craig asserted that he personally found philosophical arguments for (2) more substantial than scientific arguments for it, but given the debate topic and his opponent, he attempted to defend (2) by way of contemporary cosmology.  Craig's evidential defense of (2) fell into these two categories:

a) Evidence from the expansion of the universe
b) Evidence from the second law of thermodynamics

Obviously, the technical scientific data that might confirm or deny the plausibilities of (a) and (b) are beyond the expertise of most of this debate's audience (myself included).  However, the thrust of these two arguments are not (seemingly) too difficult to grasp conceptually.

With respect to (a), one largely only need entertain the familiar idea of the Big Bang, which is supposed to be the absolute beginning of our contiguous time-space universe.  The primary reason the Big Bang has been theorized (to the best of my knowledge) is the observation that the fabric of our space-time universe is continuously expanding, sort of like the way a balloon expands when inflated.  Since the universe is observed to expand as time goes from past to future, so the universe must likewise contract as time is reversed and runs from future to past.  Theoretically, though, there is a limit to how far something can continue to contract (a balloon, for example, can only be "crunched" so far into some minimal volume).  The smallest possible point of contraction is known as the "singularity," and by mathematically working the expansion rate in reverse, this singularity is supposed to have existed some 13.7 billion years ago--the "beginning" of our universe.  Craig's case, then, simply asserted that this state of affairs is well-supported by scientific research (including the much-discussed Borde-Guth-Vilenkin theorem) and therefore there's better reason to believe our universe had a beginning than there's reason to think it didn't didn't have a beginning.

With respect to (b), one largely only need entertain the concept of entropy.  Entropy is a bit technical, but it can be coarsely defined as the amount of disorder or disorganization in a closed system.  Low entropy means that there is little disorder in the system, whereas high entropy means there is a lot of disorder.  As an illustration, consider a mug of coffee.  Suppose a bit of cream were added to the coffee, and at the moment the cream entered the coffee, the cream was in the form of a perfect sphere.  This would be a highly organized state of affairs, so the entropy of the system would be low.  Then, after the first moment of entry, the cream begins to disperse throughout the coffee until it reaches a state of uniform distribution (or equilibrium) throughout the mug--the coffee is now everywhere a lighter shade of brown instead of black.  This final state of the system would have high entropy.  It's also worth noting that many theories of time equate the arrow of time with the direction of increasing entropy.  The future always has more entropy than the past.  With the concept of entropy in place, we can understand Craig's argument with respect to it from his opening speech:

"According to the second law of thermodynamics, entropy in a closed system almost never decreases.  Given the naturalistic assumption that the universe is a closed system, the second law implies that, given enough time, the universe will come to a state of thermodynamic heat death, whether cold or hot.  Given  that the universe will expand forever, it may never reach a state of equilibrium, but will grow increasingly cold, dark, dilute, and dead."

Craig uses the second law of thermodynamics to support the claim that the universe had a beginning.  He asserts that the initial low-entropy state of the universe is best explained by the proposition that the universe came into being with the low entropy as an initial condition.  Since the beginning of the universe 13 billion years ago, entropy has been continuously increasing across the entire universe.  Craig argues against the proposition that the universe is past eternal by asking, "Why, if the universe has existed forever, is it not now in a cold, dark, dilute, and lifeless state?"  The point again seems to me to be fairly easy to grasp: If there's been an infinite amount of time for the universe's entropy to increase, then the universe should have by now reached the stage of coldness, darkness, and lifelessness.  Since the universe clearly has not reached such a state (we are here, alive and well it seems), then it could not be the case that the universe is past eternal.

If Craig's first two premises are successfully taken to be more plausibly true than false, he has then successfully shown that the conclusion is likewise more plausibly true then false.

The Fine-Tuning Argument

Craig states his argument from fine-tuning this way in his opening speech:

"Scientists have been stunned by the discovery that the existence of intelligent, interactive life depends upon a complex and delicate balance of fundamental constants and quantities, like the gravitational constant, or the amount of entropy in the early universe, which are fine-tuned to a degree that is literally incomprehensible."

This argument suggests that the "life-permitting" range of values that many of the physical constants and quantities our world could have is extremely small compared to all values they could have had.  Some examples of constants and quantities include the strength of gravity, the strength of nuclear forces, the mass of the electron, and the like.  The fact that the range of possible life-permitting values of these constants and quantities is so narrow, and the fact that our universe is inside of that narrow range, leads to the demand for explanation.  This demand for explanation would not be in place if it turned out that the range of possible life-permitting values were extremely wide.  The computations behind this fine-tuning are again well beyond the expertise of the intended audience (myself included) but it appeared uncontroversial throughout the debate that the calculations behind the fine-tuning were reliable.  Carroll did object that the fine-tuning was purposive in the debate, though--i.e. he did not concede that the fine-tuning was for "us."  Craig went on to state his argument formally in his first speech:

1. The fine-tuning of the universe is due to either physical necessity, chance, or design.
2. It is not due to physical necessity or chance.
3. Therefore, it is due to design.

Premise (1) seems to me relatively uncontroversial (Carroll did not challenge it in the debate), though one who is a thorough-going materialist might reject "design" as an alternative simply because such a person would banish agency in all of its forms from the universe.  For those who do not banish design wholesale from their quiver of explanatory options, premise (1) appears to be acceptable.  I, at least, cannot think of any fourth alternatives.

Premise (2) is seemingly more controversial.  Dr. Craig argued that the fine-tuning cannot be due to necessity because there is a broad range of possible values the constants and quantities could have taken, and the ones that got fixed in our universe were fixed independently of any physical laws of nature.

"The laws of nature are consistent with a wide range of the values of these constants and quantities.  For example, the most promising candidate of a theory of everything to date, superstring theory or M theory, allows a cosmic landscape of around 10⁵⁰⁰ different universes governed by the present laws of nature so that it does nothing to render the observed values of the constants and quantities physically necessary."

As best as I could make out, this claim went unchallenged by Carroll, and for good reason.  There just is no sense in which some meta-physical (or supra-physical) laws could have forced our universe to have the values that it does.  In fact, to go that route would tread dangerously close to admitting a transcendent force behind the universe, and that would be a very awkward, if not self-defeating, position for the naturalist to take.

With respect to the chance alternative, Craig did not offer much argument with respect to the size of the improbability of the chance hypothesis; rather, he simply asserted that the odds of a fine-tuned universe arising by chance are simply too small to be rationally acceptable.  In order for one to posit chance as a serious hypothesis, it seems that there must be some theoretical probabilistic boundary (or rejection region) that spells out where the threshold between believability and non-believability lies.  It would be trivial to simply assert that no matter how improbable the fine-tuning, chance is automatically explanatorily adequate.  If we grant Craig's assertion that the degree of fine-tuning in our universe is "literally incomprehensible," it seems reasonable to concede that the odds in favor of chance are well below the threshold for explanatory adequacy.

Craig did go on to consider a popular defense of the chance alternative--that of the multi-verse.

"In order to rescue the alternative of chance, its proponents have therefore been forced to adopt the hypothesis that there exists a sort of world ensemble, or multi-verse, of randomly ordered universes of which our universe is but a part.  Now comes the key move.  Since observers can exist only in finely tuned worlds, of course we observe our universe to be fine-tuned."

This mult-verse argument is an attempt to resolve the significant odds against chance by inflating the number of trials the universe/multi-verse had to hit the fine-tuning "jackpot."  The reasoning is pretty simple.  If a gambler only has one pull at the slot machine, the odds of hitting the jackpot are extremely low; however, if the gambler has an infinite number of pulls at the slot machine, then a jackpot becomes essentially guaranteed.  The reasoning with the multi-verse argument appears to be parallel.  The only extra piece added to the multi-verse view is that by hitting the jackpot, our universe brought us into being.  In any universe that missed the jackpot, we wouldn't exist.  So, it's only possible for us to observe worlds where the jackpot was struck, which means there's no problem posed by the fact that we observe a finely-tuned world.

So, Craig pointed out that the multi-verse view requires these two features to avoid the fine-tuning problem:

i. The existence of a specific type of world ensemble
ii.An observer self-selection effect

Craig only mentioned (but did not specify) that objections to (i) exist, but he spent all of his time objecting to (ii).  He objected to (ii) on the basis of the Boltzmann Brain problem.  Craig unfortunately did not explain too well what this problem was for his audience, so it's likely many in the audience were perplexed by it.  However, the Boltzmann Brain problem is not too difficult to grasp.  If we suppose that universes fluctuate in and out of existence by virtue of random physical (or quantum physical) events in a way the multi-verse scenario might propose, then it would also be the case that entities of any kind could fluctuate in and out of existence by the same process.  So it need not be only universes that are infinitely multiplied by the multi-verse hypothesis, but even stand alone brains replete with first-person observers could fluctuate in and out of existence indefinitely.  It would not be necessary for "observers" in general to only exist in finely-tuned worlds, for observers could exist in non-finely tuned worlds where they might merely have the "illusion" of existing in a finely-tuned world.  So, Craig argued, for the self-selection effect piece of the multi-verse proposal to work, it would have to be the case that "ordinary" observers (i.e. those like us who exist in finely-tuned worlds) would be more prevalent throughout the multi-verse than "Botzmann Brain" observers who do not only exist in finely-tuned worlds.  But, it would seem that the prevalence of Boltzmann Brains would well outnumber the prevalence of ordinary observers because the ordinary observers are constrained by the prevalence of finely-tuned worlds.  Insofar as there are more non-finely-tuned worlds in the multi-verse than finely tuned worlds (which appears to be uncontroversial), we should expect more Boltzmann Brains to exist in the multi-verse than ordinary observers.  Hence, the self-selection effect loses its edge, and the multi-verse hypothesis cannot explain why we are ordinary observers in a finely-tuned world rather than Boltzmann Brains in a non-finely-tuned world.

With these considered rejections of necessity and chance as plausible explanations for the fine-tuning of our universe, Dr. Craig then asserted that premise (2) of the fine-tuning argument is more plausibly true than false.  He pointed out that it is still possible to argue that design is comparably as problematic as that of chance and design, but left that task to Sean Carroll.  If Craig was successful in showing premises (1) and (2) of the fine-tuning argument to be more plausibly true than false, then we should agree that the conclusion of the argument--that the fine-tuning is due to design--is more plausibly true than false.

These two arguments--the Kalam Cosmological Argument and the Fine-Tuning Argument--comprised Craig's primary case for the assertion that cosmological evidence lends support to philosophical arguments that favor theism.

Dr. Carroll's Negative Case

Recall Dr. Carroll's three main points:

1. Naturalism works
2. The evidence is against theism
3. Theism is not well-defined

I discussed (1) earlier in the essay and suggested that Carroll's defense of this claim was either ineffectively stipulative or incomplete.  Claims (2) and (3) comprise Carroll's negative case--his case against theism.

I want to list a few general comments about Carroll's two negative claims against theism, then discuss his specific objections to Craig's two arguments.  The first comment is that he did not seem to offer much empirical evidence in support of (2).  He instead appeared to rely on theoretical models as a kind of "evidence" against theism.  For the majority of Craig's arguments from cosmological evidence (like the expansion rate of the universe and increasing entropy supporting a beginning of the universe or the fine-tuning supporting the inference to design), Carroll responded to them by pointing out that there are models that can avoid the implications of Craig's arguments.  However, he did not generally offer empirical evidence contrary to the evidence Craig presented.  E.g., he did not say there is evidence that the universe is not expanding, that the entropy of the early universe was not low, or that fine-tuning does not exist.  He rather seemed to dispute Craig's theoretical interpretations of the evidence, which seems to me to be a suspicious way of defending the claim "the evidence is against theism."  The argument that came closest to being empirical from what I could gather was Carroll's claim that the entropy of the universe was much lower than it needed to be for the universe to be life-permitting, a response to Craig's fine-tuning argument.

(Note: As I mentioned in a note to my discussion of Carroll's positive case, Carroll did briefly offer a list of features of the world that he said fit better with naturalism than theism.  This portion of his presentation could likewise be considered a defense of the claim that the evidence is against theism, but again I'm going to let that part dangle due to the complications of addressing all of the features he listed.)

Following the observation that (2) was not a major part of Carroll's negative case, my second general comment is that he relied primarily on (3)--that theism is not well-defined. Carroll himself seemed to admit this much in the beginning of his second speech :

"[Dr. Craig's] disagreements with number two--the evidence is against theism--is largely based on using number three--the fact that theism is not well-defined."

This in and of itself is not exactly a problem with Carroll's case, but it does strike me as a weaker strategy than putting more emphasis on (2).  Irrespective of whether or not Craig used a vague definition of theism, whatever evidence there is against theism should, by and large, carry plenty of weight on its own. 

This point aside, though, my third general comment is that Craig's arguments did not even make mention of the word "theism" anyway, so Carroll's appeal to (3) at all in response to Craig's arguments is basically a red herring.  As Craig stated explicitly, his appeal to empirical evidence was only to support theologically neutral premises.  Specifically, Craig only used empirical evidence to support the claim that the universe began to exist and that the fine-tuning is not due to necessity or chance.  Whether or not "theism" is well-defined is irrelevant to the empirical data in support or opposition of those premises.

My final general comment is that, even if granted for the sake of argument, (3) is ineffectual because theism is not any more poorly defined than Carroll's conception of naturalism.  To make this point, let me first record how Carroll discussed (3) in his opening speech:

"The real problem is not with the definitions, it's when you connect the notion of God to the world we observe. That's where apparently an infinite amount of flexibility comes in, and I'm going to be inveighing against using that in cosmology."

Aside from this debate, "naturalism" in general comes in many varieties, just as theism does.  In fact, in the Oxford Handbook of Philosophy of Religion, Paul Draper (presently at Purdue University) defines metaphysical naturalism pretty loosely as the position that "supernatural entities do not exist" (p. 279).  In practice, this rejection of supernatural entities seems to comprise the majority of the common ground between naturalists.  But, this definition of naturalism allows for a wide array of versions of naturalism, which would seem to simply put naturalism in the same position Carroll puts theism in his third point.  Furthermore, within this debate, Carroll made frequent appeals to models (plural) to get around various obstacles.  He did not stick to one specific model to define his naturalist picture, but rather granted himself a host of mutually inconsistent models of the universe, all of which were presumably compatible with his definition of naturalism.  Now, this broad definition of naturalism would not by itself be a problem, but it becomes a problem if Dr. Carroll wants to assert claim (3).  This is so because the demand for precise definition cannot apply assymetrically to the two views under discussion.  Furthermore, Dr. Carroll accuses the theist of using his slippery definition of theism to wiggle out of various explanatory problems.  However, he seems to use the same slipperiness within naturalism to produce a wide variety of mutually inconsistent models to likewise wiggle out of various explanatory problems. 

With respect to this last point, a questioner during the Q and A portion of the debate thankfully raised essentially (though not exactly) this problem to Dr. Carroll for clarification.  Here is Dr. Carroll's full reply:²

"I think that there is a difference in principle between the theist trying to use the idea of God to explain all these different aspects of the universe and the scientist developing many, many mutually inconsistent models and not knowing which one is right until they are developed. I think that with every one of these scientific models there is an expectation, indeed a demand, that when we understand the model perfectly it will make absolutely, unambiguous, un-wriggle-out-of-able-of, predictions about what the universe is like. I think this is not in principle possible with theism....I think that theists would not claim that once we understand God perfectly we can predict the mass of the Higgs Boson. But if physicists would claim that once we have the correct theory of everything we will be able to predict the mass of the Higgs Boson. I think that's an absolutely crucial distinction."

I have several replies to Carroll's statement here.  First, although he is right to point out that there is a "difference in principle" between what theists are doing and what scientists are doing,  this difference is trivial and irrelevant to the point at issue.  Theologians do not do what scientists do, and vice versa.  If they did, then "theologian" and "scientist" would just be synonymous labels.  It is illicit to superimpose the aims of the scientist onto the aims of the theologian, then criticize the theologian for not operating parallel to those scientific aims.  As Craig stated at the beginning of his third speech: "I hope that it's been obvious tonight that I am not offering God as a theory to compete with scientific theories about the universe."  Given that Sean Carroll selected Alex Rosenberg for his panel of non-theist participants for the event, Carroll might sympathize with Rosenberg's "scientism," which Rosenberg advanced in his 2011 book The Atheist's Guide to Reality.³  Rosenberg's scientism essentially makes the sciences a monopolistic authority on all of knowledge and reality.  If Carroll is influenced by this view, that would explain why he tried to understand theism as a kind of scientific theory that should make predictions about the Higgs Boson or whatever.  But I think it should be obvious in this sort of debate that one cannot just presuppose something like scientism at the outset--doing so would beg the question against theism.

Second, it's not clear that predictions are even necessary components of valid scientific theories anyway.  If one counts historical sciences like archeology as "legitimate" sciences, predictions do not seem to play a central role in historical theorizing.  For example, if a scientist wishes to give an account of how a pyramid came into existence, what sort of demand for prediction would be required of this scientist?  Pyramids are not the kinds of things that routinely come into existence in ways amenable to predictions.  The same can be said of the origin of the universe, for example.  Presumably, the origin of our universe was a one-time thing (even under the multi-verse, we only observe our universe).  Thus, attempts to explain this one-time historical event are not obviously deficient if they fail to make precise predictions about the mass of the Higgs Boson and the like.

Third, as a consequence of confusing the nature of theism with scientific theory, Carroll's reply misses the main thrust of the questioner's point.  Namely, there appears to be a double-standard in place between his criticism of theism's pliability on the one hand, and no such criticism of naturalism's pliability on the other.  Since theism is a metaphysical proposition, its proper rival is the metaphysical proposition of naturalism (not the standards of scientific practice).  As such, if theism is poorly defined due to an abundance of explanatory escape hatches, then the abundance of such escape hatches in his naturalism ought to be subject to the same criticism.  Pointing out that naturalistic models make certain predictions whereas theism does not does nothing to alleviate this problem.  It is not the models that are in competition with theism; it is instead the metaphysical proposition of naturalism that is in competition with theism.  Thus, the duplication of many mutually inconsistent models, all falling under the banner of naturalism, shows that naturalism is extremely pliable.  Furthermore, Carroll liberally used this pliability throughout the debate to avoid the force of Craig's arguments, which is precisely the motivation for the criticism of poor definition in the first place.

So, in sum, Carroll's points (2) and (3) are rather problematic.  Point (2) received very little attention in the debate, which put the majority of emphasis for Carroll's case on (3).  However, (3) is at best irrelevant to Craig's arguments, and at worst equally problematic for naturalism.

Now, I will turn to Carroll's specific objections to Craig's two arguments.

The Kalam Cosmological Argument (Carroll's Response)

Recall Craig's Kalam Cosmological Argument:

1. If the universe began to exist, then there is a transcendent cause which brought the universe into existence.
2. The universe began to exist.
3. Therefore, there is a transcendent cause which brought the universe into being.

Carroll's first move was to reject (1), which was surprising because it is the metaphysical premise in the argument, whereas (2) is the empirically sensitive premise. In any case, he asserts that the premise is "false," but he does not so much as argue that it's false as say Craig has not sufficiently justified it.  There is a major difference between arguing that a premise is false and arguing that a premise is insufficiently justified.  To argue that a premise is false requires a strong demonstration that it is false.  In this case, it could involve demonstrating that there exists a universe that began to exist that was not brought into existence by a transcendent cause; or, more plausibly, it would involve demonstrating that some contingently existing entity (like a rabbit or something) began to exist uncaused.  In the absence of such a demonstration of the premise's falsehood, one can at best just doubt that it is true, not assert that it is false. 

Carroll goes on to say that "[Premise (1)] is not even false.  The real problem is that these are not the right vocabulary words to be using when we discuss fundamental physics and cosmology."  Carroll then points out that if you look in quantum theory or general relativity textbooks, you do not find the words "transcendent cause" in them, but rather differential equations.

Now, this objection makes evident my earlier claim that Carroll might be in the thrall of Rosenberg's scientism.  He's largely arguing that since scientists do not make use of concepts like transcendent causes when developing models with predictive power, then transcendent causes play no roles in the origin and operation of our universe.  In short, he's making the practice of scientists the central authority on what does or does not explain our universe.  Carroll's argument can be formalized as follows:

i. If a cause or entity explains any features of our universe, then it will be included in our scientific theories about the universe.
ii. Transcendent causes are not included in our scientific theories about the universe.
iii. Therefore, transcendent causes do not explain any features of our universe.

Premise (i) is a statement of scientism.  This argument therefore begs the question because the theist (and the audience) is under no obligation to presuppose that scientism is true.  Dr. Craig was offering philosophical arguments supporting theistic conclusions; he was not offering alternatives to the differential equations found in quantum theory textbooks because he does not embrace the scientism of premise (i) above. Carroll's gambit against premise (1) of the Kalam therefore falls flat because it presupposes that no transcendent causes explain anything in the world (by virtue of scientism) to show that no transcendent cause could explain the origin of the universe.

If one considers the phrase "transcendent cause" independently of the demands of scientism, one can see (easily, I think) that the phrase is coherently defined.  "Transcendent" simply means "beyond the space-time universe," and "cause" means "power to bring about effects."  The fact that the sciences do not have the power to reach beyond the space-time universe does nothing in and of itself to show that causes do not exist beyond the space-time universe.  Hence, Craig's premise (1) remains plausible despite the fact that quantum theory textbooks and cosmologists like Carroll do not use that language.

In his second speech, Carroll argues against (1) of the Kalam further.  He suggests that, within the physical universe, there are unbreakable laws of physics and an arrow of time, both of which provide the framework in which we can make sense of the concepts of cause and effect.  When we go outside of the physical world, though, he argues that we have no right to think that our concepts of cause and effect apply there as well. He states:

"The idea that our intuitions about cause and effect that we get from the everyday experience of the world in this room, should somehow be extended without modification to the fundamental nature of reality is absurd"

This strikes me as a rather sweeping assertion.  What Carroll seems to be suggesting is that we have no grounds whatever for theorizing what might or might not be true about the fundamental nature of reality.  Insofar as any of us are members of our physical universe, any proposition any of us might advance about the fundamental nature of reality, according to Carroll, is no better than the negation of that proposition.  Proposing that contingent entities fundamentally require causes is in no way more plausible than proposing that no contingent entity ever fundamentally requires a cause.  The propositions, presumably, are equally "absurd."

Now, this move by Carroll does not seem to strengthen his case against Craig because Carroll's assertion applies equally to him as it does to Craig.  At best, this move gets Carroll to global skepticism about the fundamental nature of reality; it does not get him naturalism, a multi-verse, or a refutation of the Kalam.  With respect to the latter, this move would make premise (1) of the Kalam inscrutable--it is neither more plausibly true than false, nor more plausibly false than true.  Instead, it would just be beyond evaluation in principle.  But, why should one adopt Carroll's skepticism about our ability to reason out fundamental principles of reality?  The commitment to the physical world and only the physical world for explanatory justification is a commitment of methodological naturalism or scientism, but one need not subscribe to those epistemologies.  Certainly, a theist is not going to subscribe to those views; so, in a debate against a theist, Carroll cannot just assume something like scientism at the outset without merely begging the question.  Furthermore, it is not in any way obvious to me that a priori "intuitions" are automatically unreliable indicators of fundamental truths, as Carroll's statement implies.  He himself appears to put a lot of stock in mathematics (e.g. differential equations), but mathematics is a quintessential case of a priori reasoning.  So it appears to me that Carroll's appeal to skepticism about the fundamental nature of reality is either question-begging by virtue of its reliance on a kind of scientism, or equally deadly to his own proposals about what the nature of reality is ultimately like (i.e. his assertion of naturalism).

So, Carroll's case against premise (1) of the Kalam has three primary claims: a) that Dr. Craig did not sufficiently justify the premise, b) that modern scientists do not use transcendent causes in their models of the universe, and c) that the fundamental nature of reality is unknowable in principle.  Neither of these claims appears strong enough to show that the premise is false.  They might be enough to show that the premise is more plausibly false than true if there were not good reasons to think that the premise is true.  However, there does seem to be some good reasons for thinking premise (1) is true: namely, the base metaphysical principle that nothing comes from nothing.  This principle gives coherence and intelligibility to our universe, because contingent effects are explainable by way of causes.  If one denies this principle, then some, if not all, contingent effects would be inexplicable.  The scientific enterprise that Carroll champions would be ruined by the denial of this principle.  What good is a model that makes "predictions" if effects are not in any way tied to causes, but instead free-floating events with no rhyme or reason to their occurrences? 

With respect to premise (2) in the Kalam, my best estimate of Carroll's argument was that several beginning-less models of the universe exist, so the universe most likely did not have a beginning.  This argument strikes me as ineffective in a number of ways.  First, he admits that the beginning-less models of the universe he discusses are not likely correct.  If this is so, his conclusion does not seem to follow from his premises.  If several incorrect beginning-less models of the universe exist, it does not follow from this that the universe most likely did not have a beginning.  To assert otherwise would put him in the same position he says the theist is in--that of granting oneself an infinite latitude in how he defines his position.  Second, in order to rescue his argument from being logically invalid, he would have to rely on a kind of scientism again.  He would have to propose that the reason a host of incorrect beginning-less models of the universe support the claim that the universe did not have a beginning is that these models are part of contemporary scientific cosmology.  These models are built by scientists, and, as scientism tells us, scientists are the central authorities on what is and is not likely true about our universe.  But, once again, presupposing scientism in this debate is question-begging.  Third, Carroll attempted to support his rejection of premise (2) by appeal to Alan Guth--one of the authors of the Borde-Guth-Valenkin Theorem.  In a series of still photos, he shows Guth holding signs that together read: "I don't know whether the universe had a beginning. I suspect the universe didn't have a beginning. It's very likely eternal but nobody knows." The thrust of this quote is presumably that Guth suspects that the universe did not have a beginning, but he certainly wrapped that statement in between a lot of uncertainty.  As a result, whatever conviction Guth might hold with respect to the eternality of the universe, this quote of his does not make it seem like premise (2) of the Kalam is more plausibly false than true.  Furthermore, after Craig's debates with Lawrance Kraus in Australia in 2013, Dr. Craig corresponded with Alex Valenkin (another author of the Borde-Guth-Valenkin Theorem) to clarify Valenkin's position.  In Valenkin's reply to Craig, Valenkin says, "I would say the theorem makes a plausible case that there was a beginning. But there are always caveats."  At best, these guys behind the BGV Theorem are very cautious about whether the universe had a beginning, and at worst seem to disagree with each other.  As far as premise (2) of the Kalam goes in light of these considerations, the layperson will lack the technical expertise to weigh the merits of the BGV Theorem, but it seems to me they are at least within their rational rights to entertain the suspicion that it had a beginning.  The two pieces of evidence Craig offered in support of this premise (the expansion of the universe and the low entropy of the early universe) do seem to add real credibility to the claim that the universe began to exist, at least to the layperson.  It would perhaps seem less plausible that the universe had a beginning if the universe were not expanding and entropy did not constantly progress from low to high.

So, Carroll's objections to premise (2) of the Kalam suffers from the fact that his collection of incorrect beginning-less models do not entail that the universe did not have a beginning, that he would have to presuppose scientism if they did entail that the universe did not have a beginning, and that the authors of the BGV Theorem are at best uncertain about the universe's beginning, and at worst in disagreement about it.  Given the weight of support the expansion and entropy of the universe add to premise (2), it seems reasonable to believe that one is justified in believing the universe began to exist given contemporary knowledge.

Overall, then, I come down on Craig's side with respect to the Kalam.  Both premises appear to survive Carroll's objections and thus remain more plausibly true than not.

The Fine-Tuning Argument (Carroll's Response)

Recall Dr. Craig's Fine-Tuning Argument:

1. The fine-tuning of the universe is due to either physical necessity, chance, or design.
2. It is not due to physical necessity or chance.
3. Therefore, it is due to design.

In his first speech, Dr. Carroll offered five objections to the fine tuning argument.  I will briefly summarize each one then comment on them.

(i) First, he objected to the claim that there even exists a fine-tuning problem in the first place.  This objection rested on the ambiguity inherent in defining "life."  His slide addressing this point read, "Our conventional notions of 'conditions necessary for life' are hopelessly parochial and anthropocentric."

(ii) Second, he argued that God should not need to fine-tune anything.  God is not supposed to be limited by the restrictions of the physical world so He should be able to do whatever He wants.  Furthermore, life is not a strictly physical phenomenon under theism anyway so life should not depend on specific physical parameters.  "In theism life is not purely physical; it's not purely a collection of atoms doing things like it is in naturalism.  I would think that no matter what the atoms were doing, God could still create life....The only framework in which you can honestly say that the physical parameters of the universe must take on certain values in order for life to exist is naturalism."

(iii) Third, Carroll argued that fine-tuning may only be apparent and might go away given more information.  He used the expansion rate of the universe as an example.  He said, "All but a set of measure zero of early universe cosmologies have the right expansion rate to live for a long time and allow life to exist.  I can't say that all parameters fit into that paradigm, but until we know the answer, we can't claim that they're definitely finely-tuned."

(iv) Fourth, Carroll defended the multi-verse model of the universe.  While he did not argue specifically that the multi-verse solves the fine-tuning problem, I think we can infer that he takes the multi-verse to do so.  He suggested that the multi-verse is not extravagant but "amazingly simple."  He said it makes predictions about what we should observe in our universe and said this is better than theism because theism does not make similar sorts of predictions.  He showed a graph of the density of dark matter and said theologians do not produce graphs like this.  Finally, he pointed out that the multi-verse can avoid the Boltzmann Brain problem.  He said this in his first speech:

"The multi-verse doesn't say that everything that can possibly happen happens with equal probability.....Different multi-verse models will have different ratios of ordinary observers to random observers....There are plenty of viable models where the Boltzman Brain or random fluctuations do not dominate."

(v) Fifth, Carroll argued that the world we would expect to exist under theism does not match up very well to the world we observe.  Carroll said this in his first speech:

"Theism fails as an explanation....If you thought [theism solved the fine-tuning problem], if you played the game honestly, what you would say is, 'here is the universe I expect to exist under theism, I will compare it to the data and see if it fits.  What kind of universe would we expect?'  I claim that over and over again the universe we expect matches the predictions of naturalism, not theism.  So the amount of tuning, if you thought that the physical parameters of our universe were tuned in order to allow life to exist, you would expect enough tuning but not too much.  Under naturalism, a physical mechanism could far over-tune by an incredibly large amount that has nothing to do with the existence of life.  And that is exactly what we observe.  For example, the entropy of the early universe is much, much, much, much lower than it needs to be to allow for life."

Carroll went on to list many other features of the world that he thought fit better with naturalism than with theism, but I will limit his point here to only the claim about entropy since it is most relevant to the fine-tuning argument.So, I will take Carroll's fifth argument here to say that the entropy of the early universe is far lower than it needed to be to permit life.

These five arguments seem to fully comprise Carroll's negative case against fine-tuning.  In his second and third speeches, I only caught him reasserting one or more of these points against fine-tuning rather than asserting new arguments.

My Commentary on Carroll's Response to Fine-Tuning

My first comment is to point out that Carroll did not address either of the premises in Craig's Fine-Tuning Argument.  He did not directly challenge either premise (1)--that the fine-tuning is due to either physical necessity, chance, or design--nor did he directly challenge premise (2)--that the fine-tuning is not due to physical necessity or chance.  However, some of Carroll's arguments can be taken as indirect challenges of these premises.

Carroll's arguments (i), (ii), and (iii) could be taken to be indirect challenges of premise (1) because (i) is an overt denial of the fine-tuning problem itself, (ii) makes fine-tuning irrelevant for the existence of life, and (iii) gives a reason to doubt there is fine-tuning. 

Carroll's argument (i) does not seem to be very decisive because it supposes that our definition of "life" is simply too poorly defined.  I do not know how Craig or other fine-tuning theorists might define life, but there does not seem to be any principled reason why the definition could not be made precise enough to understand what conditions might be necessary for life to exist. 

Carroll's argument (ii) suggests that God could make life exist in any physical world, so fine-tuning is not an indicator of whether God acted in a certain way.  This argument, though, misses the structure of Craig's Fine-Tuning Argument: Craig uses "eliminative induction" in his argument, which is a kind of process of elimination.  Even if it is true that God could make life exist in any kind of physical universe, it leaves open what explains why our universe has the specific configuration that it does.  If both chance and necessity are reasonably ruled out, and there is no alternative besides design, then design is justifiably inferred as the best explanation of the fine-tuning.  This would hold even if we did not have a good explanation for why God chose this design rather than another.  Furthermore, the theist can hold that even God is restricted by certain rules, rules like logic.  It might be logically impossible for God to make life exist in just any physical world, just as it might be logically impossible for God to make worlds containing round squares.  It seems to me, then, that any ambiguity over what God could or could not do does not decisively undermine Craig's Fine-Tuning Argument since his argument does not depend on affirmative statements of what God could do; it instead proceeds by way of eliminative induction. 

Carroll's argument (iii) appears to just be too speculative to carry much weight.  Its justification depends on future, unknown discoveries; however, it is illicit for one to reach into the unknown future and claim to know what will be found there.  Perhaps future findings will dispel the fine-tuning problem, and perhaps they won't.  No one has a better claim on future findings than anyone else, so I say (iii) is simply too speculative to do any damage. 

So, at best, Carroll's arguments (i) and (iii) allow one to doubt that there is a fine-tuning problem, but they do not establish that fine-tuning does not exist nor demonstrate that premise (1) is false.  Carroll's argument (ii) does not appreciate the structure of Craig's argument, and thus places an unnecessarily strong burden of proof on the theist to make claims about what God could or could not do.

What about premise (2)?  We can take Carroll's argument (iv) as an implicit challenge to premise (2) of the Fine-Tuning Argument.  In (iv), Carroll argued in favor of the multi-verse which, if true, would tremendously inflate naturalism's "probabilistic resources" to such a degree that a finely-tuned world would be guaranteed (much like an infinite number of trials at a slot machine guarantees a jackpot).  It does not seem to matter whether a multi-verse scenario explains the fine-tuning by way of "physical necessity" or "chance," so let us suppose that the multi-verse scenario, if true, falsifies premise (2). 

Now, Carroll's first defense of the multi-verse is that it makes predictions about what we should observe.  He suggests that this makes it better than a theistic view because theistic views do not make specific predictions, like predictions about the density of dark matter.  But, as I've mentioned several times now, this is a presupposition of scientism--it presupposes that valid explanations must be like scientific explanations.  Presupposing something like scientism is question-begging here since theists are under no obligation to become scientists. Moreover, scientific theories themselves do not always seem to require predictions anyway.  Pyramids, for example, do not routinely pop into existence in a way sensitive to predictions.  Scientific explanations of their historical origins, though, are no less scientific because they do not make certain predictions.  In general, the explanations of single historical events cannot be held to the same standards of prediction as theories of natural law or on-going phenomena are held.  Although we can agree with Carroll that the predictions of multi-verse models are well and good, we should not agree with him that theism is deficient because it does not do the same things as multi-verse models.  Theism's proper rival is metaphysical naturalism, not the methods and practice of scientists.

Carroll's second defense of the multi-verse was to assert that the Boltzmann Brain problem can be alleviated by certain models.  Here again, though, Carroll fell into the trap of doing what he criticized theists of doing: namely, using the flexibility of his view to explain away certain problems.  Carroll did not stick to one model and show how it--the one thing--not only avoided the Boltzmann Brain problem, but all other problems as well.  Instead, he appeared to suggest that the existence of models (plural) that avoided the Boltzmann Brain problem, but left unresolved other problems, was sufficient to show that the Boltzmann Brain problem is not really a problem.  But, surely, one does not solve a problem by merely relocating it.  In the absence of a singular, concrete model of the universe that simultaneously resolves all the major cosmological problems on the table, naturalism suffers from the same pliability problem Carroll ascribed to theism.  Carroll's proposed solution to the Boltzmann Brain problem could only resolve the problem if it did not create other problems elsewhere.  At least in this debate, he did not seem to suggest that he had a single naturalist model that solved the Boltzmann Brain problem yet did not create new problems elsewhere.  At best he could only say that there are models (plural) where Boltzmann Brains are not problematic, but this move quite clearly puts him in the same position as that of the theist where using pliability to explain-away problems is concerned.

What about Carroll's fifth argument?  Carroll's argument (v) is perhaps an indirect challenge of premise (2) in the Fine-Tuning Argument, but it is probably best considered as a stand-alone argument against theism and for naturalism.  Nevertheless, I'll consider it in light of the Fine-Tuning Argument as Carroll presented it as part of his discussion of that argument.  In (v), he suggested that our expectations about what the world would be like if naturalism were true align better with what we observe than they would if theism were true.  He did specifically mention the amount of fine-tuning of entropy necessary for life and concludes that it is over-tuned under the expectations of theism, but appropriately-tuned under the expectations of naturalism.  My criticism of Carroll's view here is similar to my criticism of his argument (ii).  Namely, he failed to appreciate the structure of the Fine-Tuning Argument.  What the Fine-Tuning Argument attempts to do is show that the degree of fine-tuning in the universe is inadequately explained by chance or necessity; therefore, it is best explained by the only remaining alternative: design.  Insofar as the low-entropy state of the early universe is rightly held to be inexplainable by chance or necessity, the logic of Craig's argument holds.  If, however, the low-entropy state of the early universe is adequately explainable by chance or necessity, then premise (2) of the argument would be false and the conclusion could be denied.  What seems to be at issue here is the ratio of all life-permitting entropy values to the full spectrum of possible entropy values, not the ratio of minimal life-permitting values to all possible life-permitting values.  If this former ratio is as incomprehensibly small as Craig suggested, then the fact that the early universe is "over-tuned" as Carroll said is of no consequence: it would still be in the extremely narrow range of life-permitting values rather than in the much more likely non-life-permitting range.  The latter ratio places a much higher burden on the theist to enumerate precisely what God's intentions were (e.g., God tuned the universe solely for our existence, or God wanted to be as efficient with resources as possible).  However, the Fine-Tuning Argument does not depend on such precise enumerations of God's intentions.  It is instead an eliminative inductive argument that rules out two possible alternatives in favor of the third.  The best way to refute such an argument is to show that one of the two eliminated choices is really better than the third.  With respect to fine-tuning, Carroll's argument that the entropy of the early universe was "over-tuned" falls short of this goal.

So it appears to me that Carroll's (iv) and (v) do not amount to strong criticisms of premise (2) of the Fine-Tuning Argument.  Argument (iv) either depends too much on a scientism-like presupposition (which would make it question-begging), or it depends on excessive pliability in the definition of naturalism (which would make it suffer in a way similar to Carroll's criticism of theism).  Argument (v) fails to address the Fine-Tuning Argument directly and instead illicitly attempts to raise the burden of proof on the theist.  Thus, I conclude that premise (2) survives Carroll's criticisms.

Overall, then, I conclude that both premises of the Fine-Tuning Argument survive Carroll's criticisms and thus remain more plausibly true than false.  Hence, I agree with Craig that design is the best explanation of the fine-tuning.

Concluding Remarks

My first remark is that, although I clearly favor Dr. Craig's position in this debate, much of what I have said in defense of his position was not explicitly said by Craig in the debate.  Craig chose to focus more closely on the specifics of certain naturalistic models of the universe (like Carroll's model) and attempted to illustrate problems with them.  While I agree that this is interesting in its own right, I find it stronger to challenge Carroll's whole reliance on model-building in the first place.  Perhaps Craig expected Carroll to exclusively defend Carroll's baby-universe model and thus Craig prepared to argue against naturalism by arguing against Carroll's model.  But, the fact that Carroll openly admitted that he did not think his own model was correct nullified the utility of Craig's strategy.  Carroll essentially gifted Craig with the frailty of Carroll's own model and thus made it unnecessary for Craig to argue against it.  Since I had the advantage of taking my time in thinking over Carroll's position, it became obvious to me that Carroll was simply helping himself to a plurality of naturalist models to explain-away various problems.  I thus think it would have been better for Craig to argue against Carroll's reliance on multiple models to get around cosmological problems.

Secondly, Craig could have drawn more attention to Carroll's superimposing of scientific ideals into philosophical argumentation.  Carroll was dangerously close to, if not outright guilty of, presupposing a kind of scientism in the debate.  This is somewhat understandable given that Carroll is first and foremost a scientist, but it it is nevertheless invalid for him to rely on a kind of scientism to argue against theism.  Craig did explicitly point out that theism is not a kind of scientific theory, which did go some distance in undermining Carroll's lenient appeals to scientific standards as THE standard of inquiry; but, I fear the audience might not have followed the point fully given the limited attention Craig gave it in the debate.  These criticisms of Craig aside, though, I do think he handled his case exceptionally well.

Thirdly, although I have pretty uniformly criticized Carroll's position in the debate, I do think he handled himself well also.  He did an admirable job of defending naturalism against a veteran debate opponent, and I am grateful that he avoided ad hominem attacks (and bringing up the slaughter of the Canaanites).  Insofar as audience members will grant him the presupposition of scientism (or something similar), his conclusions do seem to follow from that presupposition.  However, as I've argued, that presupposition is unwarranted in a debate over theism.  If I am right in supposing that Carroll sympathizes with Rosenberg's scientism (since he put Rosenberg on his panel for the event), then this quote from Rosenberg's book helps illustrate my point: "[T]his book is written mainly for those of us who are already deniers [of theism], not just doubters and agnostics."  In short, to assume scientism is to deny theism, so it is circular to depend on scientism to show that theism is false.

Fourth, thanks to Dr. Craig, Dr. Carroll, and all those who collaborated to make this event possible. 

---

<sup>1. Paul K. Moser and David Yandell offer a similar argument in their essay in Naturalism: A Critical Analysis.  See my review of their essay here.↩

2. The questioner appears at about the 2:22:00 mark in the debate video posted by tacticalfaith.↩

3. You can find my review of the book on my blog and on Amazon.↩</sup>