I’ve now adverted to Allen Orr’s piece in the New Yorker a few times on this blog. The first I learned of it was when a fact-checker for the New Yorker contacted me before the article came out. Once it came out, I noted that virtually none of the suggestions and corrections I had offered to the fact-checker made it into the final article.
As articles against intelligent design go, this one is not that bad. At least it gives some sense of the scientific issues that ID raises. But it also misrepresents ID in some key respects. Moreover, it has the fault of not moving the ball forward from Orr’s extended review of my book No Free Lunch that appeared in the Boston Review in 2002 (go here). I responded to that review at length on my designinference.com website (go here). Indeed, you’ll learn as much from reading this previous exchange as from reading Orr’s New Yorker article and my response to it here.
Nonetheless, because this is the article of the hour, let me address some of its problems, particularly as Orr addresses my own work and that of Michael Behe. I’ll go through the article bullet point fashion:
**Orr’s criticisms of Behe’s work are the same ones that he raised in his book review for the Boston Review of Darwin’s Black Box. His review came out out in Dec/Jan 1996/97 (go here) and was followed by a symposium of pro- and anti-ID advocates (go here and scroll down to “Darwin in the Details” — Behe was among the respondents).
**The one thing new here is that Orr looks to the type three secretory system (TTSS) as a possible evolutionary precursor for the bacterial flagellum. Our side has pretty well handled this objection (see Behe’s contribution to my coedited collection Debating Design with Cambridge University Press as well as my response to Ken Miller titled “Still Spinning Just Fine“). The problem with looking to the TTSS as an evolutionary precursor of the bacterial flagellum (leaving aside that the best evidence points to the TTSS “devolving” from the flagellum rather than evolving into it) is that it is so much simpler than the flagellum. Thus, in merely pointing to the TTSS as a possible evolutionary precursor, one has not offered anything like a detailed Darwinian pathway to the flagellum.
**This last point about the absence of detailed Darwinian pathways is the Achilles heel of Orr’s criticism of Behe. Orr remarks that “Behe and his followers now emphasize that, while irreducibly complex systems can in principle evolve, biologists canÃ¢â‚¬â„¢t reconstruct in convincing detail just how any such system did evolve.” To which Orr immediately adds, “What counts as a sufficiently detailed historical narrative, though, is altogether subjective.” This last point constitutes an damning admission — indeed, it gives away the store. Is Orr saying that evolutionary theory is in the business of telling historical narratives that are purely subjective. If so, how can it constitute a science? And if not, where are the detailed Darwinian pathways that could convince any unbiased bystander that the flagellum really did evolve by Darwinian means? Orr suggests that design theorists are tendentiously raising the bar of scientific evidence for Darwinism too high. But this is not the case. Without detailed, testable Darwinian pathways that produce irreducibly complex systems like the bacterial flagellum, why should anyone believe that such pathways exist at all? Note that ID theorists are not, as Orr intimates, asking for the actual history of the flagellum. That history might have taken any number of paths. Rather, we are asking for even one path — one detailed enough to assess whether the Darwinian mechanism could in fact produce the flagellum and systems of comparable complexity.
**One theme that Orr has played up is the idea of evolution working by borrowing components, folding in novel components into evolving structures, components that are dispensable initially but then become indispensable over time. As a sheer possibility, this might account for how a bacterial flagellum might emerge. But where is the evidence? Orr doesn’t present any. Instead, he describes a hypothetical scenario in which GPS devices, though for now dispensable in automobiles, might in fifty years prove indispensable as GPS systems get used to drive our cars. An interesting speculation, no doubt. But then Orr adds a completely misleading conclusion: “At that point, G.P.S. would no longer be an attractive option; it would be an essential piece of automotive technology. ItÃ¢â‚¬â„¢s important to see that this process is thoroughly Darwinian: each change might well be small and each represents an improvement.” THOROUGHLY DARWINIAN??? This is technological evolution, with each point in the process superintended by intelligence. Moreover, the changes at any stage are hardly gradual — a GPS placed into an automobile does not fall under what Darwin called “numerous, successive slight modifications.”
**Equally misleading is Orr’s point about trying to reconstruct the “complex historical process” by which “a bustling urban street” might have come about. Granted, one may not be able to tell the exact order in which the shops appeared. But one could at least tell a plausible detailed story about how it could have happened (first that greasy spoon restaurant, then the drug store to supply meds for the indigestion caused by the greasy spoon, then a Starbucks to wash away the bad taste, etc.). To reiterate, Orr does not appreciate that ID theorists are not asking for actual historical narratives. They are asking for detailed Darwinian pathways that could have produced the complex biological structure in question (should these also serve as historical narratives, fine, but that’s not and never was the issue).
**Orr’s sense of irony seems to desert him when he attempts to justify evolution. Why, whenever the task of justifying evolution gets tough, does he resort to examples that are design-laden — GPS devices and bustling urban centers? Possible answer: evolutionary biology is one big group-think in which its practitioners can no longer imagine the need to justify their theory. Thus, in the odd circumstance when they actually do have to justify their theory, they are at a loss and turn reflexively to design-laden examples.
**Let me now turn to Orr’s criticism of my own work. First off, let me address Orr’s attempt to insert a wedge between me and Behe, and from there to argue that “it’s hard to view intelligent design as a coherent movement in any but a political sense.” How are Behe and I supposed to part company? According to Orr, “Dembski believes that Darwinism is incapable of building anything interesting; Behe seems to believe that, given a cell, Darwinism might well have built you and me. ” It’s hard to imagine how Orr could have gotten things more wrong here. Behe and I are both on record that the Darwinian mechanism of natural selection and random variation (even when supplemented with other material mechanisms) is inadequate for generating irreducible complexity and most of the complexity we see in biological systems. Where Behe and I differ is whether all organisms trace their lineage back to a last universal common ancestor. Behe thinks they do. I argue that the evidence is far from conclusive. Nonetheless, I also argue that universal common ancestry can be consistent with ID. To say that “Behe seems to believe that, given a cell, Darwinism might well have built you and me” is therefore highly misleading. Behe has offered some speculations about a putative supercell that might have all the information for all organisms built into it and that might have evolved into the totality of living forms on Earth. But this sort of preformationism is anything but Darwinian. Orr knows better.
**Orr’s touting of the fecundity of Darwin’s theory in contrast to intelligent design is misleading. He writes: “Darwinism is one of the best theories in the history of science: it has produced countless important experiments (letÃ¢â‚¬â„¢s re-create a natural species in the labÃ¢â‚¬â€yes, thatÃ¢â‚¬â„¢s been done) ….” Have a look at Coyne and Orr’s book on speciation. None of the examples of speciation in the lab are convincing. What’s more, in no instance was an irreducibly complex molecular machine formed. Orr is engaging in hype. Where biology comes up with truly interesting and exciting insights, evolution is dispensable. And where evolution is indispensable, it comes up with wishful speculations that obfuscate rather than provide insight.
**Orr charges me with backpedaling on the use of the No Free Lunch (NFL) theorems in relation to evolutionary biology. Thus, he quotes me as writing, Ã¢â‚¬Å“The No Free Lunch theorems dash any hope of generating specified complexity via evolutionary algorithms,” to which he then juxtaposes the following quote from me, Ã¢â‚¬Å“I certainly never argued that the N.F.L. theorems provide a direct refutation of Darwinism.Ã¢â‚¬Â The first of these quotes is taken from chapter 4 of my book No Free Lunch (2002), the second is taken from my book The Design Revolution (2004). I submit that I did not change my tune or beat a tactical retreat in the two years between the writing of these books. My focus in these discussions of evolutionary algorithms has always been displacement. Consider the following passage later on in chapter 4 of No Free Lunch:
The essential difficulty in generating specified complexity with an evolutionary algorithm can now be stated quite simply. An evolutionary algorithm is supposed to find a target within phase space. To do this successfully, however, it needs more information than is available to a blind search. But this additional information is situated within a wider informational context (what in the last section we called the information-resource space). And locating that additional information within the wider context is no easier than locating the original target within the original phase space. Evolutionary algorithms therefore displace the problem of generating specified complexity but do not solve it. I call this the displacement problem.
Note the logic here: blind search is inadequate (no one contests this); NFL shows that randomly searching among these searches doesn’t improve the original search (this follows from the NFL theorems according to which averaging over searches is equivalent to blind search); hence to secure an effective search requires information that picks out such a search over against blind search (this is displacement); what’s more, this information is more than the information required to successfully conduct the original search (displacement again). The subtlety here, and one Orr gives no evidence of appreciating, is that there are two searches here — the original one and a higher-order one. By the way, I’ve filled in all the mathematical details here in a paper titled “Searching Large Spaces.” I referred the fact checker for the New Yorker to various relevant sections of this paper, but it’s clear that nothing was made of this. Note also that this paper, in section 7, has a novel measure-theoretic statement and proof of the No Free Lunch theorem.
**Orr tries to get around the force of displacement by remarking that the NFL theorems don’t hold in the case of co-evolution (fitness landscapes that vary over time as organisms/population agents evolve). Actually, I dealt with co-evolution in section 4.10 of my book No Free Lunch, and argued there that co-evolution does not overcome the displacement problem. In his 2002 Boston Review critique of that book, Orr claimed that I hadn’t even addressed the issue of co-evolution when I had devoted a whole section to it. More importantly, the No Free Lunch theorem that I prove in my paper “Searching Large Spaces” does in fact hold for co-evolutionary scenarios, as does the displacement theorem that I prove there. Again, I offered this information to the New Yorker fact-checker, but it simply got passed over.
**Orr now has a long history of suggesting that co-evolution is a magic bullet for evolving biological complexity. But he raises it as a mere possibility, never filling in details. What’s more, when we look to evolutionary algorithms within computer science, we find an interesting disanalogy with what Orr is hoping evolutionary processes can accomplish in actual living systems. Within evolutionary computing, fitness can change with time as circumstances change and more information becomes available. Accordingly, the objective function can, as it were, hop around. But — and this is crucial to understand within the context of evolutionary computing and stochastic optimization generally — the objective itself does not hop around. Orr is asking us to believe that co-evolution can work effectively by changing objectives (e.g., first a TTSS for pumping poisons, then a flagellum for motility), but we have no parallel to this in the field of evolutionary computing.
**Here Orr would claim that there are no objectives — or, as he put it, “organisms aren’t trying to match any ‘independently given pattern’.” But such a claim is beside the point. If organisms are evolving by Darwinian means, then the mechanisms guiding their evolution are undirected, i.e., these mechanisms are not aiming to produce eyes or other complex structures exhibiting definite functions. But even so, the issue is not whether organisms whose evolution is governed by these mechanisms are trying to match independently given patterns. The issue is whether they actually are matching independently given patterns, even if the mechanisms supposedly responsible for their evolution are not. Complex biological structures (e.g, bacterial flagella) that function as “bidirectional motor-driven propellors” exhibit an independently given pattern regardless of whether the mechanisms that Orr thinks brought them about do or don’t.
**Orr attributes the enthusiasm with which my arguments have been met to “an innumerate public that is easily impressed by a bit of mathematics.” That’s one possibility. Another is that my mathematics is giving theoretical support to intuitions that most people have for a long time harbored.
**Toward the close of his essay, Orr quotes Michael Ruse: “It is simply not the case that people take up evolution in the morning, and become atheists as an encore in the afternoon.” While it is true that there is no strict logical contradiction between the nonteleological view of evolution that Orr holds and theism, I would ask him how many of his fellow evolutionary biologists are in fact theists. As a matter of sociology (rather than logic), Will Provine got it right: “Evolution is the greatest engine of atheism ever invented.”
**Orr concludes: “Intelligent design has come this far by faith.” Let me urge a more compelling insight: “Evolution has come this far in spite of the facts.”