Unprobable Evolution

I just read Wonderful Life by Stephen Jay Gould (1989. New York: W.W. Norton & Co.). It was the first Gould book I ever tried to read--I started several years ago, but unfortunately put aside for a while. I reread it from scratch this weekend, but this time with the added benefit of having read much of Gould’s other works, including Full House and The Structure of Evolutionary Theory (although I’ve not yet finished this massive tome). There is one specific issue relating to Wonderful Life that is relevant to some claims that our evolutionist friends have made over the years: Darwinism, despite claims to the contrary, is not predictive.

Gould illustrates this in Wonderful Life by the analogy of "running life’s tape again." This effectively demonstrates that evolution can never be predictable. At the risk of being repetitive to an extreme, we will look at several of his quotes below:

I believe that the reconstructed Burgess fauna, interpreted by the theme of replaying life's tape, offers powerful support for this different view of life: any replay of the tape would lead evolution down a pathway radically different from the road actually taken. But the consequent differences in outcome do not imply that evolution is senseless, and without meaningful pattern; the divergent route of the replay would be just as interpretable, just as explainable after the fact, as the actual road. But the diversity of possible itineraries does demonstrate that eventual results cannot be predicted at the start, and none would ever occur a second time in the same way, because any pathway proceeds through thousands of improbable stages. Alter any early event, ever so slightly and without apparent importance at the time, and evolution cascades into a radically different channel. (p. 51).

Naturally, Gould remained a Darwinist despite demonstrating the problems with Darwinism (and as an interesting side note, his examination of the philosophy of Walcott—discoverer of the Burgess Shale, who tried to shoehorn everything into the prevailing evolutionary paradigm—is relevant in looking at Gould too). But notice the framework here: evolution can be explained "after the fact" but "cannot be predicted at the start." This is illustrated further by Gould:

I challenge any paleontologist to argue that he could have gone back to the Burgess seas and, without the benefit of hindsight, picked out Naraoia, Canadaspis, Aysheaia, and Sanctacaris for success, while identifying Marrella, Odaraia, Sidneyia, and Leanchoilia as ripe for the grim reaper. Wind back the tape of life, and let it play again. Would the replay ever yield anything like the history that we know? (p. 188)

Indeed, such a task is virtually impossible. So pervasive is this unpredictability that Gould rightly asks how it is possible for us to predict an alternate evolutionary sequence when we cannot even see how the sequence that did (supposedly) occur happened:

After all, we cannot even make predictions when we know the line of descent: we cannot see the mayfly in Aysheaia, or the black widow spider in Sanctacaris. How can we specify the world that different decimations would have produced? (p. 292).

Now lest we be charged with twisting Gould’s words, it is important to note that Gould is specifically addressing his argument to catastrophes, and not to "normal" evolution. (It should also be noted, however, that this firm of a distinction is not clearly stated in Wonderful Life, although it is clearly assumed and you can spot it when you know it’s there; the clearer statements that Gould is not speaking universally of evolution here are actually found in other works Gould later penned.) Thus, Gould does seek to avoid some of the charge of circular reasoning that is inherent in the tautology problem—in fact, Gould does address this specifically, although only tangentially:

Arguments that propose adaptive superiority as the basis for survival risk the classic error of circular reasoning. Survival is the phenomenon to be explained, not the proof, ipso facto, that those who survived were "better adapted" than those who died. This issue has been kicking around the courtyards of Darwinian theory for more than a century. It even has a name--the "tautology argument." Critics claim that our motto "survival of the fittest" is a meaningless tautology because fitness is defined by survival, and the definition of natural selection reduces to an empty "survival of those who survive."

Creationists have even been known to trot out this argument as a supposed disproof of evolution (Bethell, 1976; see my response in Gould, 1977)--as if more than a century of data could come crashing down through a schoolboy error in syllogistic logic. In fact, the supposed problem has an easy resolution, one that Darwin himself recognized and presented. Fitness--in this context, superior adaptation--cannot be defined after the fact by survival, but must be predictable before the challenge by an analysis of form, physiology, or behavior. (p. 236).

Ironically, just a paragraph later, Gould refutes this very claim of "easy resolution":

But if we face the Burgess fauna honestly, we must admit that we have no evidence whatsoever--not a shred--that losers in the great decimation were systematically inferior in adaptive design to those who survived. Anyone can invent a plausible story after the fact. For example, Anomalocaris, though the largest of Cambrian predators, did not come up a winner. So I could argue that its unique nutcracker jaw, incapable of closing entirely, and probably working by constriction rather than tearing apart of prey, really wasn't as adaptive as a more conventional jaw made of two pieces clamping together. Perhaps. But I must honestly face the counterfactual situation. Suppose that Anomalocaris had lived and flourished. Would I not then have been tempted to say, without any additional evidence, that Anomalocaris had survived because its unique jaw worked so well? If so, then I have no reason to identify Anomalocaris as destined for failure. I only know that this creature died--and so, eventually, do we all. (p. 236-237)

In other words, Gould here argues that Natural Selection—or more specifically "Survival of the Fittest" is not a tautology because it is predictive; yet it is impossible to predict who will survive. In short, the way I see the argument shaping up is as follows:

Opponents to Darwinism say X neutralizes Natural Selection
Gould says that X does not neutralize Natural Selection because of Y.
Gould then says that Y is not true.

Frankly, I can’t see why Y is relevant then. To be as charitable as possible, perhaps we can argue that Gould is saying Natural Selection is true for most of the time, just not when it really matters—at the time of mass extinction (which Gould labels as "decimation"). This also does fit with some of Gould’s other statements, wherein he specifically says that Natural Selection does not apply during these mass extinctions:

Groups may prevail or die for reasons that bear no relationship to the Darwinian basis of success in normal times. Even if fishes hone their adaptations to peaks of aquatic perfection, they will all die if the ponds dry up. But grubby old Buster the Lungfish, former laughingstock of the piscine priesthood, may pull through--and not because a bunion on his great-grandfather's fin warned his ancestors about an impending comet. Buster and his kin may prevail because a feature evolved long ago for a different use has fortuitously permitted survival during a sudden and unpredictable change in rules. And if we are Buster's legacy, and the result of a thousand other similarly happy accidents, how can we possibly view our mentality as inevitable, or even probable? (p. 48)

In fact, this evolution of traits that can be fortuitously used during a catastrophe is key to Gould’s theory. Later, he calls this the "different-rules model" of catastrophic evolution, saying:

The different-rules model therefore fractures the causal continuity that Darwin envisaged between reasons for success within local populations and the causes of survival and proliferation through long stretches of geological time. Hence, this model strongly promotes the role of contingency, viewed primarily as unpredictability, in evolution. If long-term success depends upon incidental aspects of features evolved for different reasons, then how could we possibly know, if we rewound life's tape to a distant past, which groups were destined for success? Their performance and evolution during our observation would not be relevant. We might base some guesses on incidental features that usually imply survival through a mass extinction, but how could we do so with any confidence? In an important sense, these crucial features don't even exist until the different rules of mass extinctions make their incidental effects important--for extreme stress may be needed to "key up" these features, and animals may never experience such conditions during normal times. And how can we know, in our rich and multifarious world, what the next episode of mass extinction, somewhere down the road, will require? Unpredictability must rule if geological longevity depends upon lucky side consequences of features evolved for other reasons. (p. 310-311)

It is important to note that, again to be as charitable as possible, Gould is arguing that Darwinism works on "normal" times, but during catastrophe, "the causal continuity that Darwin envisioned" is fractured. Indeed, Gould’s argument is that it is because Darwinism doesn’t work when it comes to catastrophe that the unpredictability of evolution comes into play. If Darwinism was the only force at work in evolution, then it would be possible to simply look at the Burgess organisms and determine which one is better suited to survival.

There are several glaring problems if we accept Gould’s theory, though. First, the unpredictability of Darwinism means that we do not have any scientific way to examine evolution. Evolution is simply narrative, which does not fit into the strict roles of scientific method. (Gould argues that we should not heed such rules at this point, but I find this to be nothing more than a case of special pleading that he would never allow a theist.) The problem is even summed up by Gould:

Historical explanations are distinct from conventional experimental results in many ways. The issue of verification by repetition does not arise because we are trying to account for uniqueness of detail that cannot, both by laws of probability and time's arrow of irreversibility, occur together again. We do not attempt to interpret the complex events of narrative by reducing them to simple consequences of natural law; historical events do not, of course, violate any general principles of matter and motion, but their occurrence lies in a realm of contingent detail. (The law of gravity tells us how an apple falls, but not why that apple fell at that moment, and why Newton happened to be sitting there, ripe for inspiration.) And the issue of prediction, a central ingredient in the stereotype, does not enter into a historical narrative. We can explain an event after it occurs, but contingency precludes its repetition, even from an identical starting point. (p. 278)

In short, if we accept Gould’s reasoning, we have literally broken the back of Darwinism. Darwinism cannot explain which organisms will ultimately survive when it matters the most, and as a result is limited in scope to only times between the great extinctions. While this accounts for a great deal of time, if we accept the geological time scale, it still does not account for why any specific species survived the Permian extinction, or even the end of the Cretaceous period when the dinosaurs were killed off. Gould’s argument is that mammals survived not due to their being better adapted, but due purely to pure luck, and as such we could not look at a mammal in situ with a dinosaur and determine which one would survive and which one would die. Darwinism had nothing to do with survival at this point; pure chance was all that was involved. Furthermore, our reconstruction of the events can only be done via unscientific just-so stories.

The second, more important problem, is that Darwin’s theory of Natural Selection doesn’t work during the space between mass extinctions either. Remember that the different-rules model Gould proposes is predicated on organisms gaining traits that serve absolutely no survival advantage while they get them, and which only later serve a purpose when a catastrophe hits. The heart of this problem is summed up in the previously quoted passage: "In an important sense, these crucial features don't even exist until the different rules of mass extinctions make their incidental effects important--for extreme stress may be needed to 'key up' these features, and animals may never experience such conditions during normal times."

In other words, animals must gain features that serve no purpose (in survivability terms), which (because science divorces itself from the realm of teleology) cannot have been introduced intentionally, and which then are not weeded out through normal evolution. But the fact of the matter is that after a short time, a species reaches stasis when it comes to variety. As Ernst Mayr points out:

With drastic selection taking place in every generation, it is legitimate to ask why evolution is normally so slow. The major reason is that owing to the hundreds or thousands of generations that have undergone preceding selection, a natural population will be close to the optimal genotype. The selection to which such a population has been exposed is normalizing or stabilizing selection. This selection eliminates all of those individuals of a population who deviate from the optimal phenotype. Such culling drastically reduces the variance in every generation. And unless there has been a major change in the environment, the optimal phenotype is most likely that of the immediately preceding generations. All the mutations of which this genotype is capable and that could lead to an improvement of this standard phenotype have already been incorporated in previous generations. Other mutations are apt to lead to a deterioration and these will be eliminated by normalizing selection. (What Evolution Is, 2001. New York: Basic Books. p. 135, italics his).

Now Mayr was, of course, a gradualist while Gould was a catastrophist. Even so, this illustrates the problem for us. Either organisms evolve slowly and not too far away from the basic phenotype, or else organisms evolve rapidly so that when a random catastrophe occurs there will be a chance for some of the organisms to survive. This tension is part of the reason Dawkins (another gradualist) so despised Gould. When we look at organisms today, they appear to be very similar to their phenotypes—suggesting the slow evolution as put forth by Mayr; but this kind of evolution is destroyed by catastrophes, where it breaks down into oblivion as demonstrated by Gould. Clearly, the two positions cannot both be correct; and equally as clearly, they both fail at crucial points.

Most evolutionists today argue amongst themselves as to which view is right: gradualism or catastrophism. Neither side seems to grasp that there is a third alternative available: neither side is right.