It's been a month and a half since my last post in this series, and recently a friend asked me why I stopped. I can think of two reasons: first, I spent the month of March teaching a graduate course for the first time; second, I'm worried about how this is going to go. I'm worried because I can see that the book is poor scholarship – Meyer is either underinformed or overcommitted to his cause – and I can see that my critique will be considered within a religious milieu that hinders straightforward criticism and analysis. Ergo, I think this might not be very pretty. It would be a lot more fun to blog about any of 15 different papers from the last two issues of Nature.
But we need to finish, partly because I'll be on a panel of critics at an event with Stephen Meyer himself in Los Angeles next month. (Not just critics: "a powerful group of credentialed critics." More later.)
Chapter 8 is called "Chance Elimination and Pattern Recognition." It deals first with the notion of chance and then with subjects that are at the very heart of design thought – the dual consideration of improbable events and the genesis of phenomena that exhibit "patterns." The chapter is pretty good, but seems to contain seeds of significant future confusion.
On page 175, Meyer begins a short consideration of the question "What Is Chance?" It's one of the better answers to the question that I've seen. He correctly notes that scientists don't usually take "chance" to imply "without cause."
Instead, they usually mean that the event in question occurred because of a combination of factors so complex that it would have been impossible to determine the exact ones responsible for what happened or to have predicted it.He concludes:
Thus, when scientists say that something happened by chance, they do not usually mean that some entity called "chance" caused something to happen. Instead, they mean that there is a process in play that produces a range of outcomes each of which has a chance or probability of occurring, including the outcome in question.This seems exactly right to me. Yes, there are some phenomena that are thought to be unpredictable in principle, meaning that the impossibility of predicting the outcome is absolute and not simply due to vast complexity or human lameness. (See this week's Nature and accompanying news stories for a fascinating account of the identification of True Randomness, arising from such a phenomenon.) But that's not what we scientists (especially biologists) mean when we say "random" or "by chance."
And I think this is important. I'm a Christian who's perfectly happy talking about random mutation. I picture mutations occurring "by chance," and I have good reasons for doing so. But this doesn't mean that I assume that the mutations had no cause, or that their occurrence is random in some grand metaphysical sense. It seems to me that a lot of the emotional appeal of ID can be neutralized by emphasizing this simple point, one that Meyer makes clearly.
Then he starts a section called "Chance as Negation." It begins:
But there is usually more to the notion of chance than that. When scientists attribute an event to chance, they also usually mean there is no good reason to think the event happened either by design or because of a known physical process that must, of necessity, generate only one possible outcome.Now wait. The second part, about generating only one possible outcome, is pretty obvious. If there's not a "range of outcomes," then sure, that's not a process that we would normally attribute to "chance." But "no good reason to think the event happened by design?" That sounds odd to me. After all, a designer can put chance occurrences to very good use. Consider the random selection of experimental subjects for assignment into experimental groups (such as treated and untreated groups). The "event" of assigning a subject (by a coin toss, say) to a group might not suggest "design," but the "event" of placing subjects into groups is being deliberately and rationally influenced. And I think we should keep that in mind. Because it's all too easy for ID enthusiasts to equate "chance" with "absence of design." I think that's a mistake.
Now, I'm not saying that Meyer makes that mistake in this chapter; although I do think he overemphasizes design as the opposite of chance, his main point involves the process by which scientists and others rule out chance. He explores the ideas that underpin hypothesis testing, as made famous by R.A. Fisher. (Anyone who's ever done a t-test is perhaps painfully aware of the principles here.) Meyer describes how William Dembski expanded on Fisher's ideas by adding the concept of pattern. According to Meyer, Fisher's methods merely identify improbable events, inferring that they were then unlikely to have occurred by chance. Dembski added the notion of pattern recognition:
And that was Dembski's key insight. It wasn't the improbability of an event alone that justified the elimination of chance. It was the match between one event and a pattern that we recognize from another.Did William Dembski really contribute a "key insight" to humanity's understanding of chance and design? I don't know. Do the world's experts on statistics and probability see the same significance in Dembski's work that Meyer sees? I don't know (though I've seen evidence to the contrary). But if Meyer is saying that "mere improbability" does not justify "the elimination of chance," he's surely right, and it would be great if more ID enthusiasts knew this.
A final section of the chapter points to looming trouble for this book. The section is called "One Final Factor: Probabilistic Resources." It contains the typical calculations that show how fantastically unlikely it is for a random process to discover a particular outcome. Meyer's example (from Dembski) involves coin flipping, but it could just as easily have been, say, the generation of a phrase from Shakespeare by random selection of letters. The idea is that X (getting 100 heads in a row) is so vastly improbable that there aren't enough "probabilistic resources" in the universe to expect it to occur randomly. Sound familiar? Richard Dawkins tackled the very same problem almost a quarter century ago, using a silly little computer program that the ID scholars still don't understand. When I see an ID fan talking about "probabilistic resources" then I expect to see one of the most common and damaging errors of the ID movement: the assumption that evolution is an inherently "random" process. Maybe Meyer will surprise me.
We'll skip the typo on page 187 (seriously: who edited this book?) and go straight to our traditional comparison of Stephen and Charles.
Here's most of one of Meyer's final paragraphs, on page 193:
My investigation of how chance hypotheses are generally evaluated suggested a specific line of inquiry and raised several questions that I would need to answer. Is the chance hypothesis for the origin of biological information a substantive hypothesis or merely a vacuous cover for ignorance? If substantive, are there reasons to affirm the adequacy of the chance hypothesis? Or are there grounds for eliminating chance from consideration, at least as the best explanation?And here's Charles Darwin, on page 193 of the Origin of Species, 6th Edition:
Mr. Mivart adduces this case, chiefly on account of the supposed difficulty of organs, namely the avicularia of the Polyzoa and the pedicellariæ of the Echinodermata, which he considers as "essentially similar," having been developed through natural selection in widely distinct divisions of the animal kingdom. But, as far as structure is concerned, I can see no similarity between tridactyle pedicellariæ and avicularia. The latter resemble somewhat more closely the chelæ or pincers of Crustaceans; and Mr. Mivart might have adduced with equal appropriateness this resemblance as a special difficulty; or even their resemblance to the head and beak of a bird. The avicularia are believed by Mr. Busk, Dr. Smitt, and Dr. Nitsche—naturalists who have carefully studied this group—to be homologous with the zooids and their cells which compose the zoophyte; the moveable lip or lid of the cell corresponding with the lower and moveable mandible of the avicularium. Mr. Busk, however, does not know of any gradations now existing between a zooid and an avicularium. It is therefore impossible to conjecture by what serviceable gradations the one could have been converted into the other: but it by no means follows from this that such gradations have not existed.To learn more about Mivart and his case, check out his On the Genesis of Species, which at least wins for Best Title.