Methinks it is sort-of like two weasels

That may include sequence deletions, which of course cannot be seeing in the current sequence. That allows yet more alternative paths, and is known to lead to the evolution of Irreducibly Complex structures.


Thank you. Just don’t ask me to do a chi-squared test!

1 Like

Nowadays we have computers to do the grunt work. :slight_smile:

1 Like

No, equally improbable.

And in the case of mutations, meaningful probabilities can be calculated.

No, improbable, viewed before the fact of the 500 events.

Now you are confusing probabilities and likelihoods. Dembski’s probability bound is just that, a probability bound, not a likelihood bound.

Let me try again, consider rolling a die, the probability of any given number coming up is 1/6. It does not matter if I pick the number before or after the roll, if I am viewing the probability of the event before the fact, the probability of the number picked coming up is still 1/6.

The probability of an event, considered before the fact, is “number of successful events” / “total number of events”, regardless of when or how the successful events are picked.

But can you quantify this? We have specific numbers for Behe’s example, there are 2 successful paths, and 1 success in about 10^20 trials. So not many alternative paths to chloroquine resistance here…

13 posts were split to a new topic: Why are We Disagreeing with ID?

There is no general quantification. It must be done separately for each individual case.
As I said above simply using the probability of the path that was followed is not sufficient - more analysis is required. Identifying the alternative paths and including them in the probability calculation is a part of that analysis.


We have no such thing. Behe quote-mined a single sentence from a review.

Again, I invite you to bet to demonstrate your faith in Behe.


You don’t know this.

How would a mutation that increases resistance NOT be selectable? Walk me through the molecular mechanisms.

No, it’s a description of your position.

Did you read it? Did you understand it? How do you know there are no others?

What about the papers that weren’t cited?

Lee, I’m asking whether you are willing to bet on Behe being correct.

What is the evidence, in terms of numbers of trials? Hint: what people say about the evidence is not the evidence.

Structures that meet that definition are easy to evolve by purely Darwinian mechanisms.

Moreover, neither you nor Behe has the slightest idea which mutations were necessary.

And he reached that conclusion by misrepresenting the evidence.

No, existing variation comes from OLD mutations. Thus, Behe’s assumption that organisms are “waiting” for new mutations is untenable.


But that can’t tell you that it didn’t evolve for the same reason that such a calculation can’t tell you that a particular set of cards can’t be drawn. Merely saying that from the perspective of the ancestral/prior state some particular outcome looks unlikely can’t tell you that the outcome you do have wasn’t, in fact, produced. From the perspective of your great grandparents 10 generations back your existence would seem miraculously remote too. And yet here you are.

So I just have to ask, what is the value or significance of doing such a calculation with respect to evolution? They give us no reason to think that any extant structure didn’t evolve.

I must conclude you have no rational basis for thinking any known biological structure or system was not produced by evolution. There is no reason to think intermediates in flagellum evolution would be nonfunctional(they’d just have functions other than cellular motility), and the a priori probabilites you calculate can’t rationally support the conclusion that X didn’t evolve in the same way they can’t support the conclusion that you aren’t the product of 10 generations of ancestors.


There are two successful paths[1] that we know about. No-one has evaluated all the other possible paths to determine whether or not they would lead to chloroquine resistance, or to something else equally useful. So you have absolutely no idea how many alternative paths there might be.

The 1 in 10^20 calculation is wrong too, for multiple reasons, including (i) it doesn’t take in to account the mutations occurring in either order, (ii) it doesn’t take into account the possibility of a second form of resistance having arisen in an organism that was already resistant.

  1. Actually two successful sets of paths. ↩︎


Very well then, improbable, the point remains. You are arguing against math (the theory of statistical inference).

A likelihood is a joint probability, which describes your calculation. I echo @swamidass’s suggestion to consider why working scientists and mathematicians, Christian and atheist, do not accept this form of argument.Could it be that it simply doesn’t work for any similar example you may care to choose?


Now it’s time to consider what you have left out of Dembski’s ideas. It’s an important one - specification. While it doesn’t go quite far enough it is still a big step in the right direction. And it does explain my examples quite well.

The basic idea is simple enough and contained in the name - it’s a description of the event, and one that covers the important details of the event while omitting everything else - and removing the “everything else” from consideration when calculating the probability. The only probability that matters is the probability of meeting the specification.

A prediction is a perfectly good specification. In fact it’s the best sort of specification, because it’s made without any hindsight.

So predicting the lottery is impressive because the prediction is a specification.

The draw itself - without the prediction - (probably) has no specification, so it is not impressive at all.

Predicting more heads than tails in a sequence of fair coin tosses is a specification but it has a high probability of being right (0.5 for an odd number of tosses, no matter how many - and never less than 0.25). So that is unimpressive, too.


Right, so two paths were identified, and all paths were open to evolution, so we’re not excluding any paths by definition. And including these paths in the calculation makes no real difference, as I said before.

How so? I need details.

Yes, I do know this, 10^20 / 2 is still on the order of 10^20.

But I meant that there are two individually non-selectable mutations on each path, that are essential to resistance.

“A minimum of two mutations sufficed for (low) CQ transport activity, and as few as four conferred full activity.” (Summers et al. here) That seems quite clear. Now there may be other paths, but they would likely be less probable than 1 in 10^20, since evolution should have found them in the more than 10^20 trials it went through.

Do you mean how do we know that there were 10^20 trials? “On the other hand, resistance to chloroquine has appeared fewer than ten times in the whole world in the past half century. Nicholas White of Mahidol University in Thailand points out that if you multiply the number of parasites in a person who is very ill with malaria times the number of people who get malaria per year times the number of years since the introduction of chloroquine, then you can estimate that the odds of a parasite developing resistance to chloroquine is roughly one in a hundred billion billion.16 In shorthand scientific notation, that’s one in 10^20.” (The Edge of Evolution, p. 57)

How so? I need more than mere assertions.

But I meant that at one point these mutations were new. And this is the pertinence of Behe’s approach: he notes what evolution actually did, with both existing variation, and with new mutations.

But I’m saying it’s improbable, not that it cannot happen.

Again you are confusing the probability of an event after the fact, with the probability of an event viewed before the fact. Roll a die, with one outcome marked as “success”, the probability viewing the event before the fact is 1/6, and after the fact is either 1 or 0.

Behe’s edge of evolution is helpful here, if two new protein-protein binding sites are the limit of what evolution can do, then producing a flagellum would require many of these. Now if there are selectable stopping-points along the way, we should see flagella being formed today–but we don’t. This leads me to believe that there are no mostly-selectable pathways to a flagellum.

But as I said above, this is the beauty of Behe’s approach, to note what evolution actually did. No possible paths were excluded, mutations can occur in any order, and other forms of resistance could have occurred in an already-resistant organism.

But as far as I understand it, the likeliest mean (of say, the binomial, or exponential or normal distributions) is just the sample mean, so we’re back to 1 in 10^20 for chloroquine resistance.

Behe is ignoring what could have happened but didn’t. His approach isn’t beautiful, it’s blather.


And I’m asking what the significance of that statement is, since it can’t rationally tell us that something didn’t evolve.

No, me explaining why it isn’t rational to argue against something having occurred by the statement that X is a priori unlikely, with an analogy to your own existence, is not me confusing anything. The analogy simply makes it more obvious why the reasoning you are employing does not support the conclusion you are seeking.

But we have no reason to think it is. The diametrically opposite is the case. You have unfortunately been misinformed about biochemistry by reading creationist literature I’m sorry to have to tell you.

In fact protein-protein binding is so ubiquitous that it is practically unavoidable by chance, so much so that we know there is selection operating to suppress and reduce the chance occurrence of protein-protein interactions in the cell.

For a nice overview of this phenomenon I recommend this excellent review article:

Should one of these protein-protein binding sites however be beneficial, simple and realistic modeling work shows it is easy to select for it instead. It is known that a considerable fraction of proteins are something like one mutation away from forming binding spots on their surface that would make them able to form large fibrils and structures. I have a thread about a year ago where I reference a few papers that detail these phenomena:

How do you know we don’t? What assay did you run, how long did you observe and where? How do you know that some prokaryote out there with some membrane-embedded structure isn’t on it’s way to evolving a flagellum-like structure?


By definition you are excluding any unidentified paths.