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:
To learn more about Mivart and his case, check out his On the Genesis of Species, which at least wins for Best Title.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.
19 comments:
From the little that I know of the literature, philosophers have been arguing about the distinction between surprising and unsurprising improbable events or observations for quite a while (e.g. in arguments concerning cosmic fine-tuning and anthropic reasoning). I don't recall seeing Dembski cited in any of those works, I'm sad to say.
I think the 'Meyer Faces His Critics' event is interesting. Do we know who else is attending? Anyone from Biologos for example? One thing about ID advocates that for me really needs to be pinned down is a clear position on common descent, especially in light of modern genome analysis. I find it hard to see how they can have much to add to a scientific discussion, or even be taken very seriously in a scientific setting, if they refuse to accept the reality of common descent and indeed spend time in futile attempts to argue against it.
Mivart is a fascinating--and in some ways tragic--case of 19th century Catholic reaction to Darwin's theory. If I recall correctly, his initial opposition to Darwin modified over some years, and much of his work ended up being condemned by the Church. I think Harry Paul discussed his case in detail in his 1979 book.
"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."
I thought a significant source of biological mutations was due to radiation which in turn is generated by quantum events. My assumption is that quantum events such as the decay of an element is not even in principle predictable although statistical distributions exist. Quantum events can thus be random in a sense that events that occur whose chain of causes are too complex for us to follow are not.
Please clarify.
Dave Wallace
Wait... "powerful" critic? When did that happen?
There are three senses of "chance" (= random) in play, and there's a whole lot of slippage across them in discussions of the role of "chance" in biological evolution. They are:
1. 'Genuine' randomness, which is a characteristic of some quantum processes. Individual events (like which atom will decay next) are in principle unpredictable. The paper and news story linked in the OP describe one such system.
2. 'Complicated' randomness, in which the causal processes that bring about some event(s) are so complicated we can't figure them out, even though the event(s) might in principle be wholly deterministic and predictable if we knew everything. As in Steve's description above, we don't think they're uncaused, but they're so complicated, and perhaps due in part to unknown causal variables, that we can't predict their occurrence beyond describing the distribution they display and using that to estimate probabilities of occurrence and properties of classes of such events.
3. 'Relational' randomness, in which two variables have zero mutual information (in the information theoretic sense). Knowing the value of one of the variables tells us nothing about the value of the other. Both variables might be wholly determined and all their causal factors known and still the mutual information between them is zero.
The sense of random (= chance) that I take to be relevant when one refers to "random" mutations in evolutionary biology is the third. The statement that "mutations are random" is a statement about the mutual information between the particular mutations that occur and properties of the selective environment, namely that the particular mutations that occur are uncorrelated with (tell us nothing about) the selective environment. (The subset of mutations that are preserved by differential reproductive success in lineages are a quite different story.)
Some mutations are also random in sense 1, where the decay of an atom of a radioactive substance emits a particle and/or radiation that zaps a bond in the DNA of a germ cell, thereby altering it. Some (many? most?) are also random in sense 2. But the important sense for evolutionary biology is sense 1.
Mutations may be (at least statistically) predictable, in the sense that we know there are mutational hot spots where mutations are more likely to occur: The distribution of mutations across loci is not uniform. But again, that tells us nothing about the selective environment: the mutual information is still zero.
In many of these discussions the sense in use shifts from one to another without notice, thereby generating a whole lot of confused verbiage.
It happened when they set up the event and before they invited "powerful" critics. Biola does this kind of thing every couple of years: set up a "debate" with "critics", get the slick, organized team of ID people up there with some (usually) disorganized hapless scientists/random people vaguely plausibly portrayed to the creationist audience as skeptics, and then discuss things outside the expertise of the participants, and then declare victory when the critics can't explain the origin of life right there on stage.
Last time (2006 I think) one of the "critics" was Antony Flew, who proceeded to sleep through the event on stage.
Make sure you watch/listen to the previous events before doing this, so you at least know what you are getting into. To quote Admiral Ackbar, "It's a trap!"
Aargh!
"But the important sense for evolutionary biology is sense 3."
RBH,
Oh, good. I started to think, "wait a sec..."
:)
Nick:
Last time (2006 I think) one of the "critics" was Antony Flew, who proceeded to sleep through the event on stage.
Hilarious!
Hey all, I don't know anything more about the event yet. I've heard from one potential panelist and will soon ask for more information. Thanks to Nick and others who've emailed me privately to express warnings and concern. I know Meyer is slick and I know the crowd is creationist. I also know how to be charming, and you already know that I know a lot more about genetics, development and evolution than Meyer does. It's not supposed to be a debate, and I'm sure not approaching it in that light, but I'll get assurances from my gracious hosts that we are on roughly the same page.
RBH, that's a fantastic summary of randomness vis a vis evolution. I'll be in touch about turning it into a real post.
Dave W., I think you are wrong to assume that radiation is a major source of mutation. Most mutations, I believe, arise from errors in replication and from other sources of damage to DNA. But even if one assumes that "quantum events" are a substantial contributor to genetic diversity, one is not committed to "grand metaphysical" randomness by which I mean the kind of randomness that somehow rules out God's will or influence.
Steve,
Let me know if you need a sycophantic hanger-on to accompany you to the event, armed with FlipHD camcorder and a large rubber chicken.
Just in case there is trouble.
;)
The analysis of the three senses of randomness I described is not original with me, but I'll be darned if I can remember when or where I encountered it. It was decades ago, perhaps in the early to mid-1970s because I have the feeling that it was sometime not too long after I started professing, but beyond that my memory is blank.
Heh, well, based on that, they're probably going to helpfully pick your side's participants for you as well...how helpful. You might get stuck with some new age reporter and some random skeptic with no biology background, each of you given your 15 minutes to try and explain all of evolution and biology and origin of life science to an audience that mostly couldn't tell you the difference between DNA and proteins.
I wouldn't be that surprised if they brought Flew back for a encore and propped him up like last time, despite his recent change in status.
Thanks for this post, and to all the commenters, especially RBH, for their work, also.
Steve,
Will there be video?
Nick, since we're quoting Star Wars, here's one for ya: "Your overconfidence is your weakness."
Nick, I haven't seen the roster of the panel but I'm told that I will likely not be the only panelist who understands biology. You might be surprised.
AMW, I don't know.
Yoda, my favorite Star Wars quote in this context: "Only a Sith deals in absolutes." While reading the ideas of ID thinkers when they talk about evolution (and especially "Darwinism") I can think of few better warnings than this one: it's very hard to find an ID argument that avoids Sith-think.
There will be video of the event and you can watch live online. Visit www.apologeticsevents.com for more details!
Dembski's argument about improbability adds nothing of interest that is new to probabilists or statisticians. He uses the concept of "specified information" which he himself attributes to Leslie Orgel.
Actually Dembski's argument works roughly like this:
# The adaptations are so good that they cannot have arisen by pure random processes (such as mutation unaccompanied by natural selection)
(Most of us would hardly disagree with this)
# So you need some process such as design or natural selection to explain them. (Again, uncontroversial)
# Dembski's Law of Conservation of Complex Specified Information shows that natural selection cannot increase adaptation, so it must be design.
This last point is completely wrong. The LCCSI is both irrelevant to that point, and it is not proven!
For a more complete discussion see my article on all this in Reports of the National Center for Science Education, 2007. I think you will find it clear.
What Dembski's design detector does do is detect design-or-natural-selection by showing that pure mutation couldn't do it.
Meyer's book does not address the (devastating) criticisms of Dembski's work, and only invokes it to argue about the origin of life. Does Meyer say that Dembski's arguments prevent natural selection from bringing about adaptation after of origin of life?
Post a Comment