Best Behe takedown *ever*

1. Evolution can be modeled in terms of a static, unchanging fitness landscape. 2. The fitness landscape is a smooth, surface made up of hills and valleys, where a local minimum or maximum in any dimension is a local minimum or maximum in all dimensions. 3. The fitness function mapping from a genome to a point of the fitness landscape is monotonically increasing.

What they mean by #3 (which I agree they expressed badly) is that evolution is being modelled as a "hill-climbing" search process, in which you start somewhere or other on the landscape and *always* move uphill. So you're liable to get stuck at the nearest local maximum.

The point MarkCC makes in response is: no, actually, there's enough "noise" in the process that sometimes you can move downhill, so if there happens to be a much better hill nearby then you can get onto it. ERV adds to this the observation that what you actually have isn't a single point on the fitness landscape, but a whole lot of them, possibly sitting at lots of different local maxima. (So if there's a really big hill somewhere around there's more chance that one of them will find it.)

Unfortunately, MarkCC's point #2 is rather broken. So far as I can tell from MarkCC's and ERV's commentary (I haven't read the book myself) Behe isn't exactly claiming that "a local minimum or maximum in any dimension is a local minimum or maximum in all dimensions". Mark elaborates on this point, claiming that when the number of dimensions of the fitness landscape is very large there aren't likely to be lots of local maxima. Unfortunately, that's just plain wrong; high-dimensional fitness landscapes *do* tend to have an enormous number of local maxima.

Not that it matters; the facts the the fitness landscape changes, the search process can move downhill sometimes, and there's a large population of searchers (possibly all on slightly different fitness landscapes, too), are enough to explode Behe's argument, at least if MarkCC's account of it isn't completely garbled.

And of course there's no reason to suppose any species is at anything better than a local maximum anyway.

I think the other point is depending on the particular mutation the distance traversed on the landscape varies, so it is perfectly possible to 'jump away' from a local maximum.

g - that's what I suspected it meant (Fisher's Fundamental Theorem and all that). Of course, that's only relevant to the mean fitness, and there is variation about that.

The inability to transverse peaks does need the infinitessimal assumption too (to stop the population jumping to another peak).

On the topology of fitness functions, one of the books in the pile I should read some time is by Sergey Gavrilets, and apparently argues that the fitness landscape is largely flay, with holes, so that there aren't peaks to be jumped between. I wonder if Behe cites Gavrilets?

Bob

And when you all run the planet, there will be no more weapons of mass destruction built by atheist scientists? (After all, Dawkins assures us most scientists are atheists.)

By the way, the book is in stores now (at least Borders), so people can sit down, enjoy some coffee, and pick Behe's arguments apart (of course you could always buy the book, but what's the fun in that?)

Sort of following this discussion superficially, not being a specialist.

One thing that has come out of deep time is that the ecology, ecosystems, and thus the fitness landscape isn't static through time. In fact it is in a constant state of major flux.

Ice ages occur, they recede, oxygen levels have been much higher at 25% and much lower at 16%, the midwestern shallow sea fills in and becomes limestone, sea levels rise and fall, continents drift, old growth forests routinely burn and reset to zero, an asteroid slams in, and on and on.

In other words the fitness landscape is in a constant and serious state of flux and chance and chaos play no small part. An organism adapted to a local optimum will eventually find itself maladapted to new conditions.

The synapsids, which IIRC were advanced mammallike reptiles were an ancient lineage that looked at the time destined to end up at the top of the heap. Then something happened at the end of the triassic and they disappeared and the dinosaurs got their chance. The current theory is the dinosaurs might be building space ships about now if the asteroid hadn't hit in Yukatan.

Fitness landscapes are ephemeral and ever changing. Lineages that persist tend to be small in size, generalists, or live in buffered environments such as the deep seas.

What I found hilarious was the comment from Jehu on UcD claiming that the "No Free Lunch therom" was Dembksi's, when all Dembski did was provide some Jello topping for them.

And watch out for ERV.

Are there any anti-ID writings, no matter how ill-conceived or mean-spirited, that PT won't endorse?

Ha-haa! Gotcha Pandas Thumb! Im really a Creationist, Jane Gene, and I made that all up! Ha-haaa!

**disappears in a poof of smoke**

:P

No really, thanks for the encouragement, and please ask all the questions you want and point out every error you think/know I made. Im a student, and corrections help me learn :)

There are some other features that are implicit in what ERV was discussing. In a fitness landscape for microorganisms, very small changes can pop you into another dimension very quickly, perhaps even in a single generation (perhaps even within a generation if lateral transfers are taking place).

These possibilities are analogous to "tunneling" in nanoscale systems in physics where electrons, even atoms, can tunnel from one state to another. Now imagine a large collection of closely spaced quantum wells in which the electrons are moving. The wave function for the electrons begins to spread out into bands if tunneling from well to well becomes possible. The probability of finding an electron in a particular state becomes a superposition of the states made possible by the presence of all the wells.

And it's not even clear in real multidimensional fitness landscapes that they are orthogonal. Maybe they could be decomposed into some kind of orthogonal set, but the fact that the landscape keeps changing in time suggests (certainly on the scale of a microorganism) that the transition from one dimension to another is extremely easy.

Getting down to the scale of a virus is approaching the domains in which nanoscale systems behave. So what ERV describes is quite familiar to a physicist.

It might be an interesting exercise to attempt a Sokal-style hoax to see what exactly PT is prepared to believe about ID. I herewith offer a prize, worth up to $200, to anyone who can pull this off and afterward reveal that it was all a hoax (the precise amount to be determined by how cleverly it is pulled off).

RevBDC: It's possible that Mark Hausam is posing as a Christian YEC in order to claim your prize. I couldn't make Christianity look dumber than he does, even if you paid me a thousand times that amount!

I'm sad. I thought after the (very thorough) refutations of his first book, Behe might simply go away. But that would be overestimating how much IDists and creationists place on proposing ideas that are scientifically sound.

ERV:

Nice article. I hope that "astrological sign" and "zodiac year" in your profile are jokes too.

Nick:

Gotta admit that Behe's own testimony at Dover was one of the top takedowns too.

RevBDC: It's possible that Mark Hausam is posing as a Christian YEC in order to claim your prize. I couldn't make Christianity look dumber than he does, even if you paid me a thousand times that amount!

Frank-- I dont know how to turn those off!!!! The second I put in my birth date, those popped up! LOL!

Look, DaveScot's expertise on recognizing science seems to be limited to such pseudo-scientific endeavors as global warming denial and Intelligent design, both mostly based on ignorance rather than an understanding of science.

His musings on global warming show a lack of depth so commonly found in arguments from IDists.

Dembski need not worry though, even though Chu's focus may be on exposing the flaws in Behe's arguments, I am sure he will continue to do the same for Dembski. No need to feel left out, just jello out D.

If one examines "ERV's" "Complete Profile" at her "blog", the third factoid listed, after her age (24) and gender (female), is "astrological sign:" ("Taurus"), which is followed by "zodiac year:" ("Boar") and then "industry:" ("science"). If astrology is so important to this individual, do you really think that she has anything worthwhile to say pertaining to science? Does it not strike you as being rather absurd that someone who (rightfully) denounces "intelligent design" apparently has no objections towards astrology?

If one examines "ERV's" "Complete Profile" at her "blog", the third factoid listed, after her age (24) and gender (female), is "astrological sign:" ("Taurus"), which is followed by "zodiac year:" ("Boar") and then "industry:" ("science"). If astrology is so important to this individual, do you really think that she has anything worthwhile to say pertaining to science? Does it not strike you as being rather absurd that someone who (rightfully) denounces "intelligent design" apparently has no objections towards astrology?

Does it not strike you as being rather absurd that someone who (rightfully) denounces "intelligent design" apparently has no objections towards astrology?

Astrology has shown far more explanatory and predictive powers than ID ever will.

Werewolf:

Read comment 181573 in this thread.

Astrology has shown far more explanatory and predictive powers than ID ever will.

Astrology has shown far more explanatory and predictive powers than ID ever will.

Recombination is another thing that can get the search out of a local maximum. That is why crossovers are used in genetic algorithms.

Anyway, saying that a computer scientist does not have the credentials to judge a biochemist's writings on search algorithms is quite an amusing argument.

"Not that it matters; the facts the the fitness landscape changes, the search process can move downhill sometimes, and there's a large population of searchers (possibly all on slightly different fitness landscapes, too), are enough to explode Behe's argument, at least if MarkCC's account of it isn't completely garbled."

Indeed.

Google simulated annealing.

Could some expert here please explain the usefulness of fitness landscapes in biology?

I read the Wiki page (http://en.wikipedia.org/wiki/Fitness_landscape) and can see how they are used in genetic algorithms and other borrowings from biology, but I cannot see how the concept adds to out knowledge of biology.

Thanks

Both Mark CC and ERV say that these are three of Behe's assumptions:

quote author="Paul Power">
Could some expert here please explain the usefulness of fitness landscapes in biology?

I read the Wiki page (http://en.wikipedia.org/wiki/Fitness_landscape) and can see how they are used in genetic algorithms and other borrowings from biology, but I cannot see how the concept adds to out knowledge of biology.

Thanks

Because Biological organisms/populations can be modeled as searchers on a fitness landscape. There is a pretty intimate connection that goes both ways between the influence of how we model evolution and the development of genetic algorithms and related evolutionary computing concepts. We do something similar with protein structure stability as well, where the fitness landscape is actually a free energy landscape and minima are where we expect to find the most stable structures for a given protein sequence.

Modeling biological populations as searchers on fitness landscapes has been around for quite some time.

Werewolf Pez -

Well, first of all, the reference to astrology is probably light-hearted and humorous. I'm sure ERV doesn't "believe in" astrology.

Second of all, your "gotcha" attempt is irrelevant. Someone could wrongly endorse one pseudoscience, yet still correctly recognize the worthlessness of ID.

Having said all of that, I have often used the example of astrology to illustrate how many mainly sincere pseudoscientific beliefs are fairly harmless, whereas ID is harmful.

1) Astrology makes claims that seem intuitively at odds with, and cannot be supported by, rational science, but it does not attempt to deny the findings of mainstream science. It merely "supplements" them. Whether Mars is the "planet of manly aggression" or not, most astrologers accept the scientific description of its orbit and physical characteristics.

2) Astrology is not associated with any efforts to violate legal rights by not only inserting itself into tax-payer funded public schools as "science", but doing so in an implicit or explicit attempt to claim that one particular sectarian position is "scientifically proven" to be the "true religion".

3) Belief in astrology often appears to be sincere and to cut across the political spectrum.

ID is far worse than astrology

Werewolf Pez -

Well, first of all, the reference to astrology is probably light-hearted and humorous. I'm sure ERV doesn't "believe in" astrology.

Second of all, your "gotcha" attempt is irrelevant. Someone could wrongly endorse one pseudoscience, yet still correctly recognize the worthlessness of ID.

Having said all of that, I have often used the example of astrology to illustrate how many mainly sincere pseudoscientific beliefs are fairly harmless, whereas ID is harmful.

1) Astrology makes claims that seem intuitively at odds with, and cannot be supported by, rational science, but it does not attempt to deny the findings of mainstream science. It merely "supplements" them. Whether Mars is the "planet of manly aggression" or not, most astrologers accept the scientific description of its orbit and physical characteristics.

2) Astrology is not associated with any efforts to violate legal rights by not only inserting itself into tax-payer funded public schools as "science", but doing so in an implicit or explicit attempt to claim that one particular sectarian position is "scientifically proven" to be the "true religion".

3) Belief in astrology often appears to be sincere and to cut across the political spectrum.

ID is far worse than astrology

The double post was a result of confusion related to extremely slow loading. Apologies. I would be delighted to have one copy deleted.

Further points in favor of astrology over ID:

4) There is positive content to astrology. It does not consist of merely that "something, somehow, is wrong with astronomy".

5) Investigation into the consequences of astrology is not discouraged. Astrology remains open to growth and change.

(By the way, just in case that there is some overly literal advocate of ID around, perhaps I should add that this is not meant to be taken all that seriously.)

Evolution can explain everything that ever happened or ever will happen.

wamba-- Im running on the assumption that Marks summary is correct. If it were not accurate, Dembski would have pointed that out (I find it difficult to believe Behe would write a book on woodworking without consulting Dembski, even more so with mathematics). So if Dembski cant say 'Ah ha! Chu-Carrolls summary is wrong! Behe really writes about X, Y, Z!' and instead chooses to say 'Chu-Carroll is MEAN! Mooooooooommieeeeeee! *SCREEEEECH!!*', I feel comfortable with that assumption.

Should Dembski choose to give a grown-up response at some point, and it turns out Mark was incorrect in his summaries, I will gladly address the new facts. However, I wont be reading Behes books in the near future. Again, if he chooses to behave like a big boy and publish a peer reviewed paper on this topic, I will happily review that (ie my response to the Sternberg ERV/epigenetics paper).

Paul-- I can tell you how I use them! Ill try to keep it brief :) Okay, Im interested in how the fitness of HIV changes over time within an infected individual. Minor problem-- There are innumerable dimensions effecting the relative fitness of the viruses Im isolating from a patient. Im only interested in one dimension: a very specific region of the gene 'env'. So one of my first tasks was peeling away all the extra dimensions so I could be sure the fitness differences I was seeing between my viruses were because of that env region, and nothing else. You cant ignore all the contributors to a fitness landscapes in experimental design.
So, once I figure out how this env region relates to replicative capacity, Im going to test another parameter, like transmissibility. Are 'more fit' viruses better at transmission? Worse? Random? I can also test similar, but different dimensions-- like whether this region of env has the same effect on fitness in PBMC, lab cell lines like TZMb-1 (do our lab experiments reflect real life?), dendritic cells, etc. Are viruses that are 'more fit' in dendritic cells lead to a more aggressive form of AIDS? What if I add patient antibodies to the infection (add another dimension to the picture)?

Honestly, I find it strange that UD is bawling over Matts math credentials. None of the examples I gave to refute Behes argument required any form of math other than 'greater than' 'less than'. But I suppose when youre generating a model instead of directly taking measurements, or maybe if you have >2 dimensions, or youre looking at all possible genotypes as opposed to a population of genotypes, things require math. *shrug*

hehe Lunch-break post-- Hope it made some sense, Paul :)

Im running on the assumption that Marks summary is correct.

Could some expert here please explain the usefulness of fitness landscapes in biology? I read the Wiki page (http://en.wikipedia.org/wiki/Fitness_landscape) and can see how they are used in genetic algorithms and other borrowings from biology, but I cannot see how the concept adds to out knowledge of biology. Thanks

— Paul Power

I'm a student, and corrections help me learn :)

Ok try this again since my last post didn't work. The usefulness is that biological organisms (or more usually populations) can be modeled as searchers on a fitness landscape, this practice has been around for some time and I believe that the concept was developed by population geneticists (most of whom were incredible statisticians) first. Mathematical models of all sorts, especially ones such as fitness landscapes with searches, have been incredibly useful through the decades and are becoming more so as computational power improves.

Astrology has come up several times in this thread. If all the superstition that has accreted around astrology over the centuries were to be peeled away, might there be a small factual basis for it?

Back in the days when we were all hunter gatherers, and after our species moved away from a specific breeding season, babies could be born at any time of the year. Could there be regularly observable differences between eg a winter baby and a spring baby? After all, their food (or their mother's food) was resticted to what was available. The growth and development environment for the 2 babies would be quite different, whether they were eating food or their mother's milk. The differences might not only be in physical development (and rate of development) but also show up in personality and intelligence traits. Add to that the interactions between 'adjacent' babies as they subsequently develop and there might be further regular & observable differences.

When people looked around for reasons to explain these differences, instead of attributing them to the timely availability of food (a repeatable cycle with irregularities) they latched onto another 'obvious' cause: the movement of stars across the sky (a repeatable cycle with fewer irregularities)?

Just a thought.

Thanks to ERV and Dan for sheeding light on how fitness landscapes are used in biology.

Can you kindly add a little more to my comprehension:

How is fitness measured - in particular how many generations are taken into account?

How is distance along the x-axis produced from the genotype - is it one mutation per x-axis unit ?

Thanks

Re: astrology.
I have seen a couple of apparently well-done studies comparing date of birth with success in sports that indicated strong connections (I'm sorry, I've no idea where I saw these). The conclusion was that those who are the oldest in their age-cohort when they enter the sport as children are more likely to succeed, which makes sense.

Thanks to ERV and Dan for sheeding light on how fitness landscapes are used in biology. Can you kindly add a little more to my comprehension: How is fitness measured - in particular how many generations are taken into account? How is distance along the x-axis produced from the genotype - is it one mutation per x-axis unit ? Thanks

— Paul Power

If astrology is so important to this individual, do you really think that she has anything worthwhile to say pertaining to science?

Does it not strike you as being rather absurd that someone who (rightfully) denounces "intelligent design" apparently has no objections towards astrology?

BTW, Moron Pez, she posted nearly 3 hours before you put up your idiotic invalid inference that her sign was calculated by the blog software from her birth date.