I was recently directed to the above video featuring notable atheist Matt Dillahunty (with whom I have had interaction in the past). In it, Dillahunty makes a number of elementary mistakes that are quite common among popular-level criticisms of intelligent design. I thought it would be worthwhile to offer a short respon
Misrepresenting Intelligent Design
His first mistake is in representing the claims of the ID community.
He claims that the “modern versions” of the design argument, “put
forth by the intelligent design movement,” essentially “attempt to
use mathematics to demonstrate that the universe and life couldn’t have come about by natural processes affected by chance andtherefore there must have been a designer.” Is this really what
proponents of ID contend? Not quite.
Statistical improbability does not and cannot demonstrate that the universe and life could not have come about by chance. Dillahunty makes this point himself later in the video. But by the same token one could also argue that the evidence used against a defendant in a court of law cannot prove beyond a shadow of a doubt that the suspect is guilty. There are always going to be alternative explanations that are logically possible. But being possible does not make an explanation reasonable. That is why the standard needed to convict a suspect is demonstration of guilt “beyond a reasonable doubt.”
The second problem with Dillahunty’s representation of the claims of ID is that design is not presented as the exclusive alternative to chance. ID proponents recognize that phenomena can be explicable by chance, necessity (i.e. law-like regularities), design, or a combination of these. We recognize that evolution by natural selection is a mechanism that involves the combination of both chance and necessity. We also do not present ID as the default alternative to chance — rather, we argue that there are positive indicators of the role of conscious deliberative agency in the origins of certain natural phenomena. We thus infer design as the best explanation for the phenomena to be explained, based on our experience of cause-and-effect relationships.
What Are the Odds?
Dillahunty argues that, when we are dealing with issues like cosmic fine-tuning and the complexity of living systems, the odds of these results being actualized by chance are incalculable because we have insufficient knowledge to reliably estimate them. He gives the illustration of rolling a “one” on a die. He notes:
If all you know is that I rolled a one, you can’t calculate the odds. We might be talking about a ten-sided die or a twelve-sided die or a twenty-sided die. Without more information, you don’t know how many outcomes there are.
But it’s worse than that. There could be multiple ones on the die, which means you don’t know how many favorable results there are either.
Is it really true that we do not have a handle on the number of favorable outcomes? Dillahunty can adopt this position if he wants, but he must recognize that it represents a fringe view in modern physics. For many of the constants and parameters of physics, physicists are in widespread agreement that there are extremely narrow life-permitting ranges. As astrophysicist Paul Davies writes in the International Journal of Astrobiology, “There is now broad agreement among physicists and cosmologists that the universe is in several respects ‘fine-tuned’ for life,” (Davies, 2003).
Taking another example, Dillahunty describes a standard 52-card deck with 13 cards in 4 different suits. After dealing all 52 cards of a randomized deck into 4 hands, Dillahunty asks, what are the odds that one person will receive all 13 spades? He calculates the number of possible hands as 635,013,559,600, and notes that only one of those is comprised of all 13 spades. Dillahunty says that, despite this staggering improbability, “we know that it’s possible and that it happens.”
Here he makes another mistake. Dillahunty has offered us no information on the probabilistic resources available. Indeed, if each one of the 10 million bridge players in the United States were to be dealt twenty bridge hands each day it would take approximately eight years, eight months and one week for such a hand to be actualized. Without some information on the probabilistic resources, it is not possible to determine whether one might reasonably expect such a result to be occasionally realized.
In the case of biological systems, we do have a handle on the available probabilistic resources. Indeed, the discipline of population genetics attempts to determine waiting times of combinations of mutations given the known mutation rates, effective population sizes, and generation turnover times. Dillahunty may try to argue that we don’t have a similar handle on the probabilistic resources with respect to the fine tuning of the laws and constants of physics, since we cannot rule out the possibility of there being a plethora of other universes, each with its own distinct set of laws and constants.
But suppose Matt and I are playing a game of cards. Round after round, I am dealt a royal flush. After twenty rounds, Matt grows suspicious that the deck has been stacked. “But Matt,” I exclaim, “you cannot rule out the possibility that there are invisible card players all over the universe — it is quite reasonable, then, that I was dealt twenty consecutive royal flushes. If enough people are playing card games out there, then one of them surely would. So why are you surprised? That’s not evidence of me cheating.”
Would that explanation really fly? The fact is that we only have evidence for one universe, and you are unjustified in arbitrarily inflating the probabilistic resources merely in order to evade an inference to design.
For an excellent review of the fine-tuning argument and the multiverse, I recommend reading Barnes (2012).
Confused About Specified Complexity
Dillahunty points out that any combination of 13 cards is equally improbable. He says, “What are the odds of getting any 13 cards?” Showing a random hand of 13 cards, he says:
What’s the odds of this other hand? It doesn’t seem that the odds are the same because that other hand is not significant to us. But the odds are the same. Replace the King of Spades with a five of hearts, and the odds remain the same. And we could create a house rule that this is now the best hand.
Here, Dillahunty demonstrates that he has failed to interact at an even cursory level with ID literature. This misunderstanding on his part also came out during my conversation with him on his TV program, The Atheist Experience. It has never been the claim of the ID community that staggering improbability indicates design. Rather, both improbability and specification are required in order to justify an inference to design. The probability has to be such that the probabilistic resources are exhausted and there must be conformity to some independently given pattern.
Yes, Dillahunty is quite correct that being dealt any hand of 13 cards is as improbable as any other. However, famed philosopher Alvin Plantinga explains in his book Where the Conflict Really Lies (pp. 196-197):
Appropriately taken, this point is perhaps right; but how is it relevant? We are playing poker; each time I deal I get four aces and one wild card; you get suspicious; I try to allay your suspicions by pointing out that my getting these cards each time I deal is no less probable than any other equally specific distribution over the relevant number of deals. Would that explanation play in Dodge City, or Tombstone?
Dillahunty maintains that “specified complexity” is a concept that lacks meaning, but I would suggest that he (and any Darwinian who takes this view) is being inconsistent on this point.
Consider the proposition of common ancestry and the arguments used to support it. Biologists recognize that the nested hierarchical distribution of shared base substitutions and mobile element inserts between different lineages cries out for an explanation beyond chance. Why? After all, the probability of exactly the same substitutions and integrations occurring in the genomes of different lineages independently by chance is equivalent to any combination of substitutions and integrations. Yet the common ancestry hypothesis is favored over the chance hypothesis because of the specificity, in addition to the improbability.
This sort of specified complexity needn’t necessarily indicate design since the specification is transmitted from some other source — in the same way that an imprint in the snow with the striking resemblance to a signpost might be explicable by a signpost having fallen and left an imprint in the snow. Would Dillahunty take me seriously if I were to argue, for instance, that the nested hierarchical distribution of endogenous retroviral placement in primate genomes was not evidence for common ancestry because the likelihood of that particular distribution is no more unlikely than any other given distribution? If Dillahunty and his comrades wish to be consistent, therefore, most of the arguments for common ancestry need to be abandoned.
What About Life?
Dillahunty asserts that calculations performed by ID proponents pertinent to the origins of life and biological complexity suffer from the same problems that, so he argues, plague arguments relating to cosmic fine-tuning. “No matter how improbable it seems,” argues Dillahunty, “they haven’t demonstrated that their supernatural explanation is even possible, let alone probable.”
Here we run into another problem with Dillahunty’s argumentation. To understand the problem, consider this example: How would one go about demonstrating that a Higgs boson is possible? How would one go about calculating the probability of a Higgs boson existing? We infer the existence of the Higgs boson by observing its effects and reaching a judgment about the best explanation of those effects. In like-manner, we infer the existence of an intelligent designer by observing certain effects that are habitually associated with conscious activity.
A second issue is that it is difficult to definitively say that a proposition entails no logical incoherence — the burden of proof thus lies with he who asserts that there is some logical incoherence.
Dillahunty takes the example of the origins of protein structures and the rarity of functional folds in combinatorial space. His two critiques are that (1) improbability does not imply impossibility (what I like to call the dumb and dumber response); and (2) the initial assumptions that led to the calculation concerning the improbability of protein folds are faulty. Since (1) has already been addressed above, let me now address (2). Dillahunty complains:
What they’re really saying is that the odds of a modern complex protein structure forming in a single trial by blind chance is 1 in 10113 or whatever their calculation is. But every single one of their assumptions seems suspect at best if not outright wrong. Why did you calculate the odds for a modern complex protein structure? Well, I think it’s because if you began by calculating the odds of a simpler protein structure that could then evolve into a more complex structure, you are accepting the theory of evolution by natural selection, and that’s something that they just won’t do.
Here, Dillahunty betrays another inconsistency in his thinking. He has just critiqued ID for allegedly positing a cause without demonstrating that “it is even possible, let alone probable.” He has now turned around and asserted that the most ancient proteins were far simpler than modern proteins — without first demonstrating that such a world is even possible, let alone probable.
There are a number of obvious scientific problems with Dillahunty’s proposal, but being a non-biologist he may be forgiven. For one thing, there is the issue of the minimum number of amino acids necessary for protein folding. It is certainly true that there are short polypeptides which possess biological activity insofar as they serve as hormones, transmitters, or regulators.
However, for a protein to have anything like catalytic activity or to serve in some structural capacity, it needs to be sufficiently large to fold into its three-dimensional structure. There are very few enzymes with a size smaller than one hundred amino acids. For sure, alpha helices and beta sheets can be formed from relatively short polypeptides (e.g., an alpha helix can be formed from a sequence of ten amino acids) — but such structural arrangements arise only in the context of a larger protein structure, not in isolation. One estimate of the minimum number of amino acids necessary for the stabilization of a folded protein structure is seventy (see p. 346 of Jack Kyte’s Structure in Protein Chemistry, second edition, 2007).
Another point that may be noted is that the amino acids that contribute to the active site are typically scattered throughout the sequence. If modern proteins evolved from shorter polypeptides (in which the crucial amino acids are necessarily close together), this observation that the crucial amino acids in modern proteins are frequently not grouped together is surprising. Why? Because moving them about during their later evolution would require restructuring of the protein — a feat that I would suggest is also prohibitively improbable.
Dillahunty’s argumentation may be appealing to those who are unacquainted with ID or the classical arguments for the existence of God — which I suspect comprises the majority of his Atheist Experience audience. But to any fair-minded individual, Dillahunty’s arguments reveal that he has not interacted in any meaningful fashion with the primary literature by advocates of ID or theism.